Diagnosis of West Nile virus infections: Evaluation of different laboratory methods

Abstract

BACKGROUND

The diagnosis of West Nile virus (WNV) is challenging due to short-term and low-level viremia, flavivirus cross-reactivity, and long immunoglobulin M (IgM) persistence.

AIM

To evaluate different methods for WNV detection [reverse transcription-polymerase chain reaction (RT-PCR), IgM/IgG antibodies, IgG avidity] in serum, cerebrospinal fluid (CSF), and urine samples of patients with confirmed WNV infection.

METHODS

The study included patients with confirmed WNV neuroinvasive infection (n = 62), asymptomatic WNV seropositive individuals (n = 22), and individuals with false-positive WNV IgM antibodies (n = 30). WNV RNA was detected using RT-PCR. A commercial ELISA was used to detect WNV IgM/IgG antibodies with confirmation of cross-reactive samples using a virus neutralization test (VNT). IgG-positive samples were tested for IgG avidity.

RESULTS

The WNV-RNA detection rates were significantly higher in the urine (54.5%)/serum (46.4%) than in CSF (32.2%). According to the sampling time, the WNV-RNA detection rates in urine collected within 7 days/8-14/≥ 15 days were 29.4/66.6/62.5% (P = 0.042). However, these differences were not observed in the CSF. The median RT-PCR cycle threshold values were significantly lower in urine (32.5, IQR = 28-34) than in CSF (34.5, IQR = 33-36). The frequency of positive WNV IgM and IgG significantly differed according to the sampling time in serum but not in CSF. Positive IgM/IgG antibodies were detected in 84.3/9.3% of serum samples collected within 7 days, 100/71.1% of samples collected 8-14, and 100% samples collected after ≥ 15 days. Recent WNV infection was confirmed by low/borderline avidity index (AI) in 13.6% of asymptomatic individuals. A correlation between ELISA and AI was strong negative for IgM and strong positive for IgG. No significant correlation between ELISA IgG and VNT was found.

CONCLUSION

The frequency of WNV RNA and antibody detection depends on the sampling time and type of clinical samples. IgG avidity could differentiate recent WNV infections from long-persisting IgM antibodies.

Key Words: West Nile virus; Reverse transcription-polymerase chain reaction; Serology; IgG avidity; Cross-reactivity

Core Tip: We analyzed different diagnostic methods in patients with West Nile virus (WNV) neuroinvasive disease and asymptomatic seropositive individuals. The WNV RNA detection rate was significantly higher in the urine/serum than in cerebrospinal fluid (CSF). The RT-PCR cycle threshold (Ct) values were significantly lower in urine than in CSF and serum samples. The frequency of WNV RNA and IgM/IgG antibody detection rates depends on the sampling time and type of clinical samples (CSF or serum). The correlation between ELISA and IgG avidity was negative for IgM and positive for IgG. No correlation was observed between ELISA IgG and virus neutralization test.

INTRODUCTION

West Nile virus (WNV) is an emerging flavivirus of public health importance. In nature, the WNV transmission cycle includes birds as virus reservoirs and mosquitoes, mainly of the genus Culex as vectors. WNV is endemic in many parts of the world (Europe, the United States, the Middle East, Asia, and Africa) causing outbreaks and sporadic infections in both endemic and non-endemic areas[1]. Although the majority of WNV infections (80%) are asymptomatic or present as a non-specific febrile disease (WNV fever), some patients develop neuroinvasive disease (meningitis, encephalitis, myelitis). The mortality rate can reach 10% in severe neuroinvasive forms of the disease[2].

WNV shares many features with some other neuroinvasive arboviruses, especially flaviviruses such as tick-borne encephalitis virus (TBEV), and Usutu virus (USUV). Due to overlapping geographic distributions and clinical symptoms with these infections, the WNV diagnosis should be confirmed using virological methods[3]. The Centers for Disease Control and Prevention/European Center for Disease Control and Prevention (ECDC) laboratory criteria for WNV confirmation include: (1) Direct detection of virus by culture, antigen, or viral RNA; (2) A four-fold rise in antibody titers between acute- and convalescent-phase serum samples; or (3) Virus-specific immunoglobulin M (IgM) confirmed by detection of neutralizing antibodies. The detection of specific IgM antibodies in the cerebrospinal fluid (CSF) and the lack of IgM to other endemic arboviruses are also criteria for confirmation of WNV neuroinvasive disease (NID)[4,5].

Because the majority of arboviruses, including WNV, have a short period of replication (short-term and low-level viremia), limiting the utility of molecular tests, serology is commonly used for disease confirmation. However, antibody cross-reactivity between viruses within the same serocomplex and persistent antibodies from previous arboviral infections complicate the interpretation of serology results[6].

ELISA and indirect immunofluorescence assay are the most commonly used screening tests for WNV serology. Although next-generation tests have improved performance by selecting the best epitopes and enhancing antigen purity, challenges in serological diagnosis still exist. In samples with cross-reactive antibodies, neutralization tests such as virus neutralization test (VNT) and plaque reduction neutralization test (PRNT), which still represent gold standard serology tests, are necessary for confirmation of the first-line serology[7].

Long IgM persistence is frequently observed in WNV infections. In a study conducted in Greece, WNV IgM antibodies were detectable for more than 3 years in 12% of patients with WNV infection[8]. Moreover, a study from Huston has found WNV IgM persistence (positive or equivocal results) in 42%, 34%, and 23% of WNV-infected patients approximately 1 year, 6 years, and 8 years post-infection, respectively[9]. Long-term persistence of WNV IgM antibodies in the CSF was also observed. In a study among patients with WNV NID from Michigan, IgM antibodies were detectable in the CSF of three patients for 110, 141, and 199 days post-acute phase[10]. In cases with long IgM persistence, IgG avidity determination could be useful for the differentiation of acute vs previous WNV infection[11].

Molecular methods such as RT-PCR in blood/serum and CSF samples are limited to the minority of patients who present with ongoing viremia or central nervous system replication[7]. Recent investigations indicated that WNV is excreted in urine longer and at higher levels compared to blood[12].

WNV grows and plaques efficiently on Vero cell culture, usually inducing a cytopathic effect[13]. Virus isolation is time-consuming and laborious[14]. Because WNV is classified as a biosafety level 3 (BSL-3) agent, virus cultivation is restricted to reference laboratories[15].

In Croatia, human WNV infections (WNV NID and WNV fever) have been reported continuously since 2012 in continental counties. In addition, some other flaviviruses are endemic in the same geographic areas, such as TBEV or occur sporadically (USUV)[16].

Because there are still many challenges in flavivirus diagnostics, this study aimed to analyze the characteristics of different methods for the diagnosis of WNV infection.

MATERIALS AND METHODS

Patients

This retrospective study included patients with NID (n = 182) and asymptomatic individuals (n = 352) tested for WNV during the two Croatian outbreaks (2017-2018). Patients with confirmed WNV NID (IgM/IgG positive and/or RT-PCR positive; n = 62), asymptomatic WNV IgM and/or IgG seropositive individuals (n = 22), and individuals with false positive WNV IgM antibodies (n = 30) were analyzed (Figure 1). Serum and CSF samples were collected in all patients, while urine samples were available for 55 patients. The samples were collected in the period from 4-19 days after disease onset (median 9, IQR = 7-13). In addition, paired serum samples taken 2-3 weeks after the first one were collected in 34 patients. Characteristics of patients with WNV NID are presented in Table 1.

Figure 1 Selection of patients with West Nile virus infection and clinical samples. Samples in gray shadowed boxes were analyzed in the study. CSF: Cerebrospinal fluid; IgG: Immunoglobulin G; IgM: Immunoglobulin M; NID: Neuroinvasive disease; RT-PCR: Reverse transcription-polymerase chain reaction; TBEV: Tick-borne encephalitis virus; USUV: Usutu virus; WNV: West Nile virus.

Table 1 Characteristics of patients with West Nile virus infection included in the study.

Characteristic	Subcategory	n (%)1
Sex	Male	35 (56.4)
	Female	27 (43.6)
Age in median years (IQR)		68 (58-76)
Clinical diagnosis	Febrile headache	2 (3.2)
	Meningitis	36 (58.1)
	Meningoencephalitis	20 (32.3)
	Meningoencephalomyelitis	2 (3.2)
	Polyradiculoneuritis	2 (3.2)

¹Unless otherwise indicated. IQR: Interquartile range.

Methods

Serum, CSF, and urine samples were tested for the presence of WNV RNA using real-time RT-PCR as described previously[17]. Initial serological screening (WNV IgM and IgG antibodies) was performed using a commercial ELISA. Samples were also screened for potential cross-reactivity with other flaviviruses endemic in Croatia (TBEV/USUV). WNV cross-reactive samples were confirmed using a VNT[18]. In addition, WNV IgM/IgG positive and IgM negative/IgG positive samples were further tested for IgG avidity to confirm or exclude recent infection[11]. Characteristics of laboratory methods used for the diagnosis of WNV infection are presented in Table 2.

Table 2 Laboratory tests used for the diagnosis of West Nile virus infection.

Method	Manufacturer/Protocol	Reference values
RT-PCR	FP: AAG TTG AGT AGA CGG TGC TG; RP: AGA CGG TTC TGA GGG CTT AC; Probe: FAM-CAA CCC CAG GAG GAC TGG-TAMRA
IgM ELISA	Euroimmun, Lübeck, Germany	Ratio < 0.8 negative; 0.8-1.1 borderline; > 0.1 positive
IgG ELISA	Euroimmun, Lübeck, Germany	RU/mL < 16 negative, 16-22 borderline; > 22 positive
IgG avidity	Euroimmun, Lübeck, Germany	AI < 40% low; 40%-60% borderline; > 60% high
VNT	In house	Titer ≥ 10 positive

AI: Avidity index; ELISA: enzyme immunoassay; IgG: Immunoglobulin G; IgM: Immunoglobulin M; RT-PCR: Reverse transcription-polymerase chain reaction; RU: Relative units; VNT: Virus neutralization test.

Statistical analysis

WNV positive detection rates were presented as numbers and percentages with 95% confidence intervals (CI). The differences in the positive WNV RNA and antibody detection rates according to clinical sample and sampling time were compared using a χ² test. A Kruskal-Wallis test was used to compare the differences in RT-PCR cycle threshold (Ct) values. The correlation between ELISA IgM/IgG levels and avidity indices (AI) and IgG/VNT titer was calculated using Spearman’s rank correlation coefficient. P value < 0.05 was considered statistically significant. For statistical analysis, a Social Science Statistics program was used (https://www.socscistatistics.com/tests/).

RESULTS

WNV RNA detection in serum, CSF, and urine

The WNV-positive detection rate was significantly higher (χ² test P = 0.049) in the urine (30/55; 54.5%) and serum (13/28; 46.4%) than in CSF samples (20/62; 32.2%). Nucleotide sequencing was successful in 15 (27.3%) urine samples (Figure 2).

Figure 2 Detection rate of West Nile virus RNA by reverse transcription-polymerase chain reaction. CSF: Cerebrospinal fluid; N: Total sample number.

The RT-PCR positive detection rate depended on the sampling time. In the CSF, the positive detection rate was higher in samples collected within 7 days after disease onset (36.4%) compared to samples collected 8-14 days (31.2%) and ≥ 15 days (25.0%). In contrast, the frequency of WNV RNA positive detection was lower in in urine samples collected within 7 days (29.5%) than in urine samples collected within 8-14 days (66.6%) and ≥ 15 days (62.5%). The observed differences were statistically significant for urine (χ² test P = 0.042) but not for CSF (χ² test P = 0.828) (Table 3).

Table 3 Reverse transcription-polymerase chain reaction positive detection rates in cerebrospinal fluid and urine samples, n (%).

Days after symptoms onset	WNV RT-PCR CSF		WNV RT-PCR Urine
	Tested	Positive	Tested	Positive
≤ 7	22 (35.5)	8 (36.4)	17 (30.9)	5 (29.4)
8–14	32 (51.6)	10 (31.2)	30 (54.6)	20 (66.6)
≥ 15	8 (12.9)	2 (25.0)	8 (14.5)	5 (62.5)
Total	62 (100)	20 (32.2)	55 (100)	30 (54.4)

Data are n (%). CSF: Cerebrospinal fluid; WNV: West Nile virus; RT-PCR: Reverse transcription-polymerase chain reaction.

The RT-PCR Ct values were significantly lower in urine samples (median 32.5, IQR = 28-34) than in CSF (median 34.5, IQR = 33-36) or serum samples (median 35, IQR = 34-36), (Kruskal-Wallis test P = 0.047, Figure 3).

Figure 3 West Nile virus reverse transcription-polymerase chain reaction cycle threshold values in different clinical samples. Plots represent medians with interquartile ranges, inner and outlier points. CSF: Cerebrospinal fluid; Ct: Cycle threshold.

WNV IgM and IgG detection in serum and CSF

WNV antibody detection rates in serum and CSF samples are presented in Figure 4 and Table 4. Using ELISA, IgM antibodies were detected in 90/95 (94.7%) of serum and 44/62 (70.9%) of CSF samples. IgG antibodies were detected in 55 (57.8%) of serum and 19/62 (30.6%) of CSF samples (Figure 4).

Figure 4 West Nile virus immunoglobulin M and immunoglobulin G detection rates in serum and cerebrospinal fluid. CSF: Cerebrospinal fluid; Ig: Immunoglobulin; N: Total sample number.

Table 4 Positive detection rate of West Nile virus antibodies in serum samples according to sampling time, n (%).

Days after symptoms onset	Tested	WNV IgM antibodies		WNV IgG antibodies
		Positive	Borderline	Positive	Borderline
≤ 7	32 (33.7)	27 (84.3)	0 (0)	3 (9.3)	2 (6.2)
8 – 14	38 (40.0)	38 (100)	0 (0)	27 (71.1)	5 (13.1)
≥ 15	25 (26.3)	25 (100)	0 (0)	25 (100)	0 (0)
Total	95 (100)	90 (94.7)	0 (0)	55 (57.8)	7 (7.3)

Data are n (%). IgG: Immunoglobulin G; IgM: Immunoglobulin M; WNV: West Nile virus.

The frequency of positive results significantly differed according to sampling time in serum, but not in CSF samples. In serum samples collected within the 7 days after disease onset, the IgM-positive detection rate was 84.3%, compared to 100% in samples collected after more than 8 days (χ² test P = 0.005). Similarly, the IgG-positive detection rate was lowest in samples collected in the first 7 days (9.3% positive and 6.2% borderline). In samples collected 8-14 days after disease onset, IgG antibodies were positive in 71.1% and borderline in 13.1% of samples. All samples collected ≥ 15 days showed WNV IgG antibodies (χ² test P < 0.001) (Table 4).

The positive IgM detection rates in the CSF samples collected within 7 days, 8-14 days, and ≥ 15 days were 62.5, 73.9, and 100%, respectively (P = 0.145). The frequency of IgG detection was similar in the period ≤ 7 days and 8-14 days (28.1 and 26.1%, respectively), while it was higher in samples collected ≥ 15 days (57.1%; χ² test P = 0.001) (Table 5).

Table 5 Positive detection rate of West Nile virus antibodies in cerebrospinal fluid samples according to sampling time, n (%).

		WNV IgM antibodies		WNV IgG antibodies
Days after symptoms onset	Tested	Positive	Borderline	Positive	Borderline
≤ 7	32 (51.6)	20 (62.5)	1 (3.1)	9 (28.1)	2 (6.2)
8–14	23 (37.1)	17 (73.9)	0 (0)	6 (26.1)	1 (4.3)
≥ 15	7 (11.3)	7 (100)	0 (0)	4 (57.1)	0 (0)
Total	62 (100)	44 (70.9)	1 (1.6)	19 (30.6)	3 (4.8)

Data are n (%). IgG: Immunoglobulin G; IgM: Immunoglobulin M; WNV: West Nile virus.

Samples with false positive WNV IgM antibodies (n = 30) were also included in the study. Analyzing the IgM antibody levels, patients with confirmed WNV infection showed a significantly higher IgM (median ratio 3.8, IQR = 3.1-4.3), than patients with false-positive IgM antibodies (median ratio 1.4, IQR = 1.1-1.8) (Kruskall-Wallis test P < 0.001) (Figure 5).

Figure 5 Immunoglobulin M antibody levels in patients with confirmed West Nile virus infection and false positive samples. Plots represent medians with interquartile ranges, inner and outlier points. IgM: Immunoglobulin M; WNV: West Nile virus.

Cross-reactive WNV IgM and IgG antibodies with TBEV/USUV were detected in 4.5% and 38.2% serum samples (Figure 6). The frequency of cross-reactive IgM antibodies was significantly lower than the frequency of cross-reactive IgG antibodies (χ² test P = 0.001).

Figure 6 Cross-reactive flavivirus immunoglobulin M immunoglobulin G antibodies detected in patients with West Nile virus infections. IgG: Immunoglobulin G; IgM: Immunoglobulin M; NT: Not tested; TBEV: Tick-borne encephalitis virus; USUV: Usutu virus; WNV: West Nile virus.

WNV IgG avidity determination

IgG avidity was performed in IgM/IgG positive patients with WNV NID (n = 52) and asymptomatic WNV seropositive individuals (n = 22). In patients with NID, 50 samples showed low AI (median 13%, IQR = 7-21), and two samples showed borderline AI (42 and 40%), respectively.

In asymptomatic WNV seropositive individuals, IgG avidity was performed in IgM-positive/IgG-positive samples (n = 10) and IgM-negative/IgG-positive samples (n = 12) (Table 6). Three samples (13.6%) showed low/borderline AI indicating acute/recent WNV infection. Using IgG avidity determination, recent WNV infections were confirmed in two IgM-positive patients and one IgM-negative individual.

Table 6 Immunoglobulin G avidity in asymptomatic West Nile virus seropositive individuals.

Serology result	IgG avidity index
	Low	Borderline	High
IgM/IgG positive, n = 10	1 (10.0)	1 (10.0)	8 (80.0)
IgM negative/IgG positive, n = 12	1 (8.3)	0 (0)	11 (91.7)
Total, n = 22	2 (9.1)	1 (4.5)	19 (84.6)

Data are n (%). IgG: Immunoglobulin G; IgM: Immunoglobulin M.

A strong negative correlation (Spearman's rho coefficient -0.512, P < 0.001) between the ELISA IgM levels (ratio) and AI (%) was observed. Samples with the highest IgM levels were associated with the lowest AI values. For IgG antibodies, a correlation between the IgG levels (RU/mL) and AI (%) was a strong positive (Spearman's rho coefficient 0.802, P < 0.001) (Figure 7).

Figure 7 Correlation of immunoglobulin G avidity. A: Immunoglobulin M antibody levels; B: Immunoglobulin G antibody levels. Blue dotted lines represent trendlines. IgG: Immunoglobulin G; IgM: Immunoglobulin M.

A total of 26 IgM and/or IgG-positive samples were confirmed using VNT. No significant correlation between ELISA IgG levels (RU/mL) and VNT titer was found (Spearman's rho coefficient 0.221, P = 0.265) (Figure 8).

Figure 8 Correlation of immunoglobulin G antibody levels (ELISA) and NT titer (Virus neutralization test). Blue dotted line represents a trendline.

DISCUSSION

The diagnosis of WNV in humans is challenging due to the short arbovirus viremic phase, low-level viremia (humans are dead-end hosts), the cross-reactivity between flaviviruses, and long-term IgM persistence[6,8,11].

In the present study, WNV RNA positive detection rates depend on the sample type (46.4% in serum, 32.2% in CSF, and 54.4% in urine samples). Similarly, in a study from Italy, patients with WNV NID and WNV fever had detectable WNV RNA in urine at a higher rate (43.8%) than WNV RNA in plasma (37.5%) and CSF (7.1%)[19]. Furthermore, in a study conducted in Israel, whole blood samples were most frequently positive using RT-PCR (86.8%), followed by urine samples (58.3%). The positivity of serum and plasma was 26% and 20%, respectively, while CSF was RT-PCR positive only in 16.5% of patients. All samples were collected on average 11 days after symptom onset. These results suggested that in patients with acute WNV fever, WNV RNA is present in whole blood significantly more frequently than any other sample type tested[20].

Although the period in which WNV RNA is detectable in the CSF varies, it can last several weeks in some patients[1]. In our study, the frequency of WNV RNA detection differs according to the sampling time. In the period within 7 days after symptom onset, 36.4% of CSF samples were RT-PCR positive compared to 31.2% 8-14 days and 25.0% ≥ 15 days after symptom onset. In contrast, the frequency of positive WNV RNA detection rate in urine was lowest in samples collected within 7 days (29.4%) compared to 66.6% and 62.5% for samples collected afterward. In a recently published study from Serbia, WNV infection was confirmed by positive WNV RT-PCR in serum and/or CSF samples in 46.3% of patients. Thirty-one percent more cases were confirmed using urine WNV RNA testing. In contrast to our results, there was no association between the sampling time and WNV RNA urine positivity[12].

Consistent with our findings (the lowest median RT-PCR Ct in urine 32.5, compared to serum 35, and CSF 34.5), in a study from Israel, the WNV viral load in urine was higher than that of whole blood, CSF, serum, and plasma, even though the urine sensitivity was lower than that of whole blood[20].

WNV-specific antibody testing is currently the most widely used approach for the diagnosis of WNV infection. WNV IgM and IgG antibodies are typically detectable by days 4 and 8 after the onset of clinical symptoms[21]. Serological testing of patients included in the present study showed IgM and IgG antibodies in 94.7% and 57.8% of serum samples, respectively. In addition, 70.9% of CSF samples were IgM positive and 30.6% were IgG positive. In the early acute phase of the illness, there is a negative serological window period of detection, as the antibodies have not yet been developed[14]. Our study showed a negative IgM serology in 15.7% of serum and 37.5% of CSF samples collected within 7 days after disease onset. No false negative serum samples were detected after day 8. However, analyzing the CSF serology, 26.1% of CSF samples tested IgM negative in the period 8-14 days, whereas all samples were positive after 14 days. The IgG positivity in serum was 9.3, 71.1, and 100% in samples collected ≤ 7 days, 8-14 days, and ≥ 15 days. In the CSF samples, IgG antibody detection rates were 28.1 and 26.1% in samples collected by day 14 and 57.1% in samples collected after 14 days.

Serological cross-reactivity between flaviviruses is common due to antigenic similarities, especially between the viruses that belong to the same serocomplex. Both species-specific and cross-reactive antibodies are produced during flavivirus infections[22]. People who live in areas where different arboviruses are endemic gradually accumulate cross-reactive antibodies from previous exposures with increasing age[6]. Therefore, it can be challenging to identify the most recent infection in patients with multiple exposures to different flaviviruses[23]. Cross-reacting antibodies are frequently observed in ELISA, while some degree of cross-reactivity was also found in a more specific VNT[24]. Given that WNV and USUV are antigenically closer by genomic phylogeny than TBEV, cross-reactions between WNV and USUV in VNT are usually more common[25]. In our study, ELISA revealed WNV cross-reactivity with TBEV and USUV. IgM cross-reactivity was significantly lower (4.5% of serum samples) than IgG cross-reactivity (38.2% of serum samples). WNV IgG antibodies cross-reacted with TBEV in 35.3% of samples and USUV in 2.9% of samples.

Despite their obvious clinical relevance, IgM antibodies experience false-positive results, which can result in WNV misdiagnosis[26]. Thirty samples with false positive WNV IgM antibodies were also analyzed in the present study. Comparing the levels of false positive IgM antibodies and WNV-specific antibodies, significant differences were observed. WNV-specific antibody levels were significantly higher (median ratio 3.8) than false positive IgM levels (median ratio 1.4).

The diagnostic implications of serum WNV IgM persistence are noteworthy since the presence of IgM antibodies is generally considered evidence of an acute or recent WNV infection. Since the long-term persistence of IgM antibodies is frequently observed in patients with WNV infection[8-10], this should be taken into account when interpreting serology results. IgG avidity was useful to differentiate between recent and previous WNV infection in both patients with NID and WNV fever[11]. In the present study, using avidity determination, 80.0% of tested IgM/IgG-positive asymptomatic individuals were classified as having recent WNV infection, while 20.0% were classified as having previous WNV exposure (long-persisting IgM antibodies). Furthermore, 8.3% of IgM negative/IgG positive individuals had low IgG avidity, suggestive of recent infection.

Our study found a strong negative correlation between IgM levels and IgG avidity, consistent with previous findings in Croatian patients with confirmed WNV infection. The samples with the highest WNV IgM levels had the lowest AI values, indicating that determination of IgG avidity is not required in cases with very high IgM results[11].

Because of its high specificity, neutralization tests are the most reliable serological assays. A correlation between neutralizing flavivirus antibody detection and the presence of specific IgG in blood specimens was observed in some studies[27]. In this study, VNT was used for confirmation of cross-reactive serum samples. However, we found no correlation between WNV-neutralizing antibody titers with binding antibody levels (ELISA). Similarly, some studies have found that the ELISA titers in convalescent WNV patients did not correlate with neutralization nor did neutralization titers increase over time[28].

A limitation of this study that needs to be addressed is the small number of patients tested. Therefore, further studies on a large cohort of WNV patients are needed to confirm our observations.

CONCLUSION

The frequency of WNV RNA and antibody detection depend on the sampling time and type of clinical samples, which should be considered when interpreting the results. Testing should include flaviviruses that circulate in a specific area to exclude cross-reactive antibodies. IgG avidity determination is a useful diagnostic method for differentiating IgM positivity in recent infections from long-persisting WNV IgM antibodies.