Seroprevalence studies suggest that a proportion of infections by TOSV are asymptomatic or paucisymptomatic. Additional studies will be necessary to evaluate the ratio of symptomatic versus asymptomatic or paucisymptomatic infections.

In some cases, TOSV infection causes a self-limiting febrile illness without CNS manifestations; these patients are not usually hospitalized, and their cases are not usually investigated further. This fact may account for the probable underestimation of TOSV infection rates.

After an incubation period ranging from a few days to 2 weeks, disease onset is intense (70%) with headache (100%, 18 h–5 days), fever (76%–97%), nausea and vomiting (67%–88%), and myalgias (18%). Physical examination may show neck rigidity (53%–95%), Kernig signs (87%), poor levels of consciousness (12%), tremors (2.6%), paresis (1.7%), and nystagmus (5.2%) (L. Nicoletti, pers. comm.). In most cases reported so far, CSF contained >5–10 cells with normoglycorachia and normoproteinorachia. Blood samples may show leukocytosis (29%) or leukopenia (6%). The mean duration of the disease is 7 days, and the outcome is usually favorable.

Although TOSV infection in most cases consists of a mild disease with a favorable outcome, a small number of severe cases have been reported in the literature. Two young brothers and a sister living in Umbria experienced TOSV infection in the form of severe meningoencephalitis with stiff neck, deep coma, maculopapular rash, diffuse lymphadenopathy, hepatosplenomegaly, renal involvement, skin rash with lamellar desquamation, a tendency to bleed, and diffuse intravascular coagulopathy. CNS manifestations occurred after 3 weeks of fever. Convalescence was marked by hydrocephalus that required a ventriculoatrial shunt. Diagnosis was established by serologic means and by PCR sequencing (32). Two cases of encephalitis without meningitis were recently diagnosed by serologic testing and by detecting TOSV sequences in CSF (33). One case of meningitis, complicated by abducens nerve palsy, was reported (34). To date, no published data exist that suggest that TOSV causes any other manifestations. However, a substantial proportion of infection likely results in asymptomatic or paucisymptomatic cases (5,35).

Seroconversion and the detection of IgG, IgM, or both, can be achieved in cells infected with TOSV. However, cross-reactivity exists between members of the genus Phlebovirus and specifically between TOSV and other serotypes of sandfly fever Naples virus.

ELISAs have been developed with either crude antigens or purified virus obtained from infected cells. The advantage of ELISA resides in its capacity to rapidly test a large number of specimens; however, cross-reactions most likely will be observed. Recently, an ELISA test based on a recombinant nucleoprotein gene was developed and is now available commercially from an Italian company. Recent seroprevalence studies were based on this test (8,21).

PRNT is the test of choice when the virus species must be confirmed. Therefore, seroprevalence data must be carefully interpreted since in most cases, analyses were performed with ELISA or IFA that cannot discriminate between sandfly fever Naples virus, sandfly fever Sicilian virus, and TOSV.

Viruses can be isolated from clinical samples by using CSF but not serum. CSF specimens that yield virus through cell culture are collected in the first 2–4 days of the disease.

TOSV replicates in a variety of animals. Intracranial, intraperitoneal, and subcutaneous routes lead to death in newborn mice, and intracranial and intraperitoneal routes lead to death in weanling mice. This effect is seen with viruses from only a few families, including flaviviruses, which are also implicated in viral encephalitis. In guinea pigs and rabbits, intracranial injection results in paralysis and death, whereas intraperitoneal injection is not fatal and results in antibody synthesis.

TOSV replicates in Vero, BHK-21, CV-1, and SW13 cells with cytopathic effect and not in C6/36 cells. However, cell culture appears to have a low sensitivity for detecting TOSV since only 14% of the PCR-positive CSF specimens added to Vero cells led to viral isolation.

In some cases, the relatively low level of virus in blood and CSF samples hampers attempts to isolate the virus. In such cases, molecular techniques based on PCR are more sensitive than IgM detection or viral isolation. Three different methods for molecular diagnosis of TOSV have been developed (Table). To date, all studies aimed at the molecular detection of TOSV sequences in the CSF of patients with meningitis or other CNS manifestations have used classic PCR detection through single-round or nested protocols. Tests of an RT-PCR assay alone, without a further nested PCR step, showed that this method did not appear to be valid for detecting TOSV, since no sample was positive after the first reaction. Two systems (14,36) use specific primers in the S segment, and the other is based on degenerate oligonucleotides targeting of the L segment (37). The most widely used has been successful for TOSV diagnosis in Italy (13) and France (R. Charrel et al. unpub. data). In 2003, a new method for detecting TOSV of Italian or Spanish origin was produced by using degenerate primers. The description of 2 genotypes of TOSV demonstrates a need for caution when designing molecular methods for diagnosis to avoid false-negative results. Recently, real-time PCR systems, including a fluorescent dye–labeled probe, have dramatically reduced the risk of contamination. The sensitivity of real time RT-PCR is close to that obtained by nested PCR protocols, and the results are obtained within 3 hours. However, to develop real-time PCR assays that detect all variants of TOSV circulating in Mediterranean countries and causing diseases in humans, a considerable amount of work must be done to determine the sequences of strains reflecting viral heterogeneity observed in different countries. The recent report of a Spanish genotype, genetically divergent from the strains circulating in Italy, which is not detected by PCR systems previously reported in Italy, underlines the requirement for a large program of strain isolation and full-length genome sequencing to achieve this goal.

The prototype TOSV strain, ISS Phl.3, isolated from P. perniciosus in 1971 has been completely sequenced. A total of 84 virus strains were obtained from 16,374 male and female sandflies (P. perniciosus and P. perfiliewi) collected in 2 localities of the Tuscany region of Italy between 1980 and 1985. Thirty-seven (44%) were identified as TOSV and 47 (56%) as a new member of the phlebotomus fever serogroup, Arbia virus. The overall virus isolation rate from sandflies was 0.5%. Viral isolation rates for both viruses were similar in different years and in the 2 localities, suggesting that the 2 virus types were active in the sandfly population simultaneously (maximum activity in July) (3). Seventeen strains of TOSV have been isolated in Spain from patient specimens (22). Several strains have been isolated in southeastern France from patients with clinical cases and remain to be characterized.

A number of strains from Italy have been partially sequenced, and only minor differences in the nucleoprotein were found among strains isolated in the early 1980s from both species of sandflies, from the bat, and from humans, with <1 amino acid substitution (L. Nicoletti, pers. comm.). Similar results were described in a study on some variants in the N gene of strains isolated from humans from 1995 to 1998; only 1 variant showed a single amino acid substitution of an 80-amino-acid region (38). Changes in the amino acid sequence that make this protein less efficient in its interaction with the viral nucleic acid may kill the virus.

A different situation has been described in Spain for partial sequences in the large segment encoding the polymerase activity. A phylogenetic analysis performed from L segment sequences obtained from 11 clinical isolates from Granada and compared with the homologous sequence of an Italian reference strain showed that Spanish sequences were closely related to one another and distantly related to the Italian strain (37). This finding suggests the presence of at least 2 geographically distinct populations of TOSV.

To date, sequence data are too scarce to perform significant phylogenetic analyses. We must therefore set up a large program of complete genome sequencing of the strains collected in different regions and simultaneously to encourage the development of viral isolation programs in all countries surrounding the Mediterranean where vectors are circulating to better understand the genetic diversity, phylogenetic relationships, and mechanisms driving the evolution of TOSV (Figure 2).

The virus has been isolated from P. perfiliewi and P. perniciosus, the most abundant sandfly species present in Italy. However, other vector species, found in different geographic areas, could transmit the virus. Serologic results indicate that the virus is present in many areas of the Mediterranean basin. Entomologic studies must be conducted to better understand the distribution and identification of potential vectors of TOSV.

The recent introduction of West Nile virus into North America has stimulated a renewed interest among health authorities regarding arthropodborne viruses, specifically concerning human blood products. Until 2002, the risk of transmitting West Nile virus to a naïve patient from a blood donation was considered negligible, given the supposed short time (≈6 days) and low viremia titers. However, ≈30 cases of viral transmission were documented in 2002 and 2003 in the United States and Canada, as well as cases of West Nile virus infections after organ transplantation from a viremic donor. Moreover, 540 positive blood donation samples were detected by using PCR, which underlines the necessity of this kind of test in an epidemiologic situation similar to that seen in the United States. Recent data on TOSV circulation in Mediterranean countries during the summer raise concerns about potential implications for blood donations.

Limited studies have been conducted on the genetic variability of TOSV. The work of Sanchez-Seco on the L segment demonstrated the presence of 2 geographically distinct populations of the virus (37). However, the study was performed on strains isolated from patients with acute neurologic disease. On the basis of seroprevalence in a healthy population, Magurano and Nicoletti hypothesized that among the different strains of TOSV that may circulate in the same area and infect humans, only a few cause severe disease, whereas most strains induce antibody response with minor or no symptoms of illness (39). The role of different strains in the symptoms and influence on the severity of TOSV infection requires clarification.