Review of Detection and Quantification of Rabies Virus Antibodies

Abstract

Rabies is an almost invariably fatal disease. According to the World Health Organization (WHO), rabies virus neutralizing antibody (RVNA) titers of ≥0.5 IU/mL are considered adequate for rabies protection. Therefore, detection and quantification of RABV antibodies are important. Many methods have been developed for detecting RABV antibodies. In the present study, we reviewed several methods of detecting RABV antibodies in human and animal samples and evaluated and compared their performance. Of 34 methods, 5 demonstrated unsatisfactory sensitivity or specificity. The others exhibited sensitivity and specificity of ≥75%. The correlation coefficient for five of eight methods was >0.8. The Bland–Altman mean bias of five of five methods was <±2.0. The kappa values of 25 of 28 methods were higher than 0.4, demonstrating at least moderate agreement. Analysis of the performance of these methods emphasized that any new technology should be considered carefully and objectively before being used as an appropriate and applicable alternative.

Introduction

Rabies is recognized as a high-priority zoonotic disease caused by an infection with rabies virus (RABV) of the genus Lyssavirus and family Rhabdoviridae (53). It kills an estimated 59,000 people annually worldwide, remaining a major public health threat in Asia, Africa, India, and many developing countries (21,26). Typical symptoms of rabies are aggression, hydrophobia, anemophobia, progressive paralysis, and hypersensitivity to sound, light, wind, and pain; some rabies patients also display abnormal sexual behaviors (68). Although the mortality rate due to rabies is very high, standardized postexposure prophylaxis (PEP) treatment can effectively prevent the occurrence of rabies. A complete course of human rabies vaccine is part of the PEP protocol (56). After completing the vaccine course, the body produces rabies virus-neutralizing antibodies (RVNAs) (80 –82). Usually, immunization is serologically evaluated 1–3 weeks after the last injection. According to the World Health Organization (WHO), RVNA titers of ≥0.5 IU/mL are considered adequate for rabies protection (79,81,89). Therefore, evaluation of RABV antibody titers is of paramount importance.

At present, researchers have developed a variety of methods to detect RABV antibodies. Serological methods for detection of RABV-specific antibodies include antigen-binding methods, antibody function tests, and antigen function assays. Conventional methods include antigen-binding assays for detecting immunoglobulins that include both neutralizing and non-neutralizing antibodies, such as the enzyme-linked immunosorbent assay (ELISA) (49) and indirect fluorescent antibody (IFA) (67) test, and antibody function assays for detecting neutralizing antibodies, such as the rapid fluorescent focus inhibition test (RFFIT) (64) and fluorescent antibody virus neutralization (FAVN) test (10). In antibody function assays, antibodies are quantitated by the live virus that escapes neutralization. RFFIT and FAVN test are currently the gold standard for detecting RVNAs (52,82). They can effectively determine whether rabies vaccines have induced neutralizing antibodies in humans and animals (28,46). However, the FAVN test and RFFIT require the use of live RABV, expensive antirabies fluorochrome antibody conjugate, and a biosafety level 2 (BSL-2) containment facility (9,50,59). In other words, special laboratory infrastructure and stricter biosafety protections are needed to perform the assays. In an antigen-binding assay, the antigen or antibody is fixed to a solid surface, such as a tube, plate, or bead. Rabies antibodies in cerebrospinal fluid or serum are characterized by their ability to bind to various antigens directly. Whether the antigen and antibody are bound and the quantity of binding are determined using a visualized detection system. Antigen-binding assays are suitable for large-scale immunization surveillance in rabies prevention.

This study reviewed some detection methods of RABV-specific antibodies. Performance values of these methods, including sensitivity, specificity, and agreement with the gold standard, were quantified.

Different Methods for Detecting RABV Antibodies

Mouse neutralization test

The mouse neutralization test (MNT) is an in vivo method (77,22). It is a complex time-consuming assay. Furthermore, in vivo antibody assays that utilize death of the host as an endpoint are inherently imprecise. Many uncontrollable factors, such as nonspecific deaths and susceptibility or resistance to lethal infections, should be considered (39). In 1973, Smith et al. (64) indicated that the RFFIT was a satisfactory substitute for MNT. Presently, however, MNT is still the officially recommended method for estimation of rabies-neutralizing antibodies (59).

FAVN test

The FAVN test is a well-established test recommended by the World Organisation for Animal Health (OIE) and the WHO. It is most commonly used to evaluate RVNAs in animal serum, but it can also be used for human serum (86). Cliquet et al. (10) indicated that distinguishing negative serum samples from positive serum samples with low titers is much easier with the FAVN test than with the RFFIT. One previous study modified the FAVN test by replacing the fluorescein isothiocyanate (FITC) conjugate with a peroxidase conjugate and then using an automatic multichannel spectrophotometer to make the result more automatable and documentable (30).

Rapid fluorescent focus inhibition test

RFFIT is an efficient standard assay recommended by the WHO, OIE, and Advisory Committee on Immunization Practices (ACIP) for determining levels of RVNAs in human serum samples. It involves a shorter test duration of 2 days (compared with 14 days for the MNT). This is a clear benefit for patients. RFFIT is complex. A number of factors, including the quality of cells and virus, variability of assay reagents, and skill and expertise of analysts, can affect its performance. Timiryasova et al. (69) identified and evaluated conditions that affected RFFIT performance and RVNA detection by evaluating parameters of the assay. Compared with the FAVN test, RFFIT is somewhat subjective with respect to plate reading. It requires laboratory operators to read the percentages of fluorescent foci in the wells; it is not “all or nothing” as in the FAVN test. Furthermore, the conventional RFFIT still uses live RABV as a challenge virus. To circumvent these limitations, some laboratories use enhanced green fluorescent protein (eGFP) to make plate reading more quantitative and efficient and to eliminate the requirement of BSL-2 laboratory containment. In this assay, eGFP is inserted into the RABV genome or other pseudotype viruses. The recombined virus is used as a challenge virus in subsequent experiments (5,33,43,44,57,66). It is combined with array scanning or ultraviolet microscopy to analyze the expression and quantification of RABV-infected cells.

Under normal conditions, RFFIT requires a volume of 50 μL of serum per test (14). Kuzmin et al. (35) developed a microneutralization test that only requires 3.5 μL of serum. Instead of using 8-well or 96-well slides, the microneutralization test uses 4-well (6-mm) Teflon-coated glass slides, therefore decreasing the reagent content by ∼90% and representing cost savings compared with the conventional RFFIT (65). Furthermore, plate reading of the microneutralization test differs from that of conventional RFFIT. In the microneutralization test, 10 separate fields are counted for each well. If a reduction or absence of fluorescence is observed, the serum sample is subjected to additional titration, in dilutions from 1:10 to 1:1,250. Only samples that have a 50% endpoint neutralization titer >1 log10 (e.g., fewer than five fields contain infected cells at a serum dilution of 1:10) are considered positive (35). This method solves the problem of the minimum volume of small laboratory animals that may be difficult to obtain for a standard RFFIT.

IFA test

The IFA test is a type of antigen-binding assay that measures a standard antigen–antibody reaction. It is one of the simplest and most rapid tests for detection of antibodies against RABV in serum and cerebrospinal fluid (27,38,67). This method detects RABV-specific IgG and IgM antibodies. Previous studies documented that the IFA test detected antibodies earlier than RFFIT (4,25,42), and in some cases, the result of the IFA assay was positive, while that of the virus neutralization assay was negative (8,24). The opposite was rare. The IFA test largely detects the RABV ribonucleoprotein instead of neutralizing antibodies; therefore, this test is rarely used as a measure of protection following vaccination (7,42).

Enzyme-linked immunosorbent assay

ELISA is widely used to detect molecules such as antibodies, antigens, proteins, glycoproteins, and hormones (1). It is simple, cheap, and accessible and its results are less influenced by the quality of the sample (12). Although some commercial kits have poor correlation with RFFIT for testing samples (13,60), ELISAs are nowadays considered suitable tests for quick and easy detection of rabies antibodies in large-scale screening programs (73). There are four classic types of ELISAs for detecting rabies antibodies: direct ELISA, indirect ELISA, sandwich ELISA, and competitive ELISA (13,34,48,49,75,76,83). Furthermore, Zhao et al. (87) successfully developed a Chinese hamster ovarian cell line stably expressing the RABV glycoprotein as the capture antigen. This ELISA kit shows high specificity and sensitivity with RFFIT. Additionally, Realegeno et al. (58) developed a new recombinant RABV nucleoprotein-based indirect ELISA kit that can detect either IgM or IgG antibodies generated against the RABV nucleoprotein. It can detect a recent or primary infection earlier.

Immunochromatographic test

The immunochromatographic assay is a technique in which a cellulose membrane is used as the carrier and a colloidal gold-labeled antigen or antibody is used as the tracer (84,85). The immunochromatographic test strip (ICTS) makes detection easier, faster, cheaper, and more convenient. This test does not require preincubation of the antigen and antibody; it only requires adding serum to a sample pad. If the virus-neutralizing antibody concentration is ≥0.5 IU/mL, the test strip will produce bands at the test line and control line. If a sample has a concentration of <0.5 IU/mL, the test strip will only produce one band at the control line (70).

The rapid neutralizing antibody test (RAPINA) utilizes this principle of immunochromatography. Compared with RFFIT and ELISA, RAPINA was established later. In this method, test serum is preincubated with an optimal amount of inactivated RABV--(iRABV) and then added to the sample hole. If the level of RVNAs in test serum is inadequate (<0.5 IU/mL), unbound iRABV will bind with the gold-conjugated monoclonal antibody (mAb) against glycoprotein (G) and the antigen–antibody complex will be trapped in the test line. If the level of RVNAs is adequate, no band will form in the test line (63). However, this kind of protocol sometimes misclassifies RVNA values >0.5 IU/mL as <0.5 IU/mL. Nishizono et al. (51) improved RAPINA. They replaced the gold-conjugated mAb against G with the virus-neutralizing mAb, AD-8. The iRABV G specifically binds with the virus-neutralizing mAb, AD-8, to make the result more accurate. ICTS and RAPINA are very suitable for prescreening large numbers of serum samples out in the wild.

Latex agglutination test and magnetic protein microbead-aided indirect fluoroimmunoassay

In the latex agglutination test (LAT), latex beads are coated with purified rabies glycoprotein. Then, the test serum is incubated with the coated beads. If the concentration of rabies-specific antibodies is adequate, agglutination is clearly visible to the naked eye. However, as shown in a previous study, the detection limit is 2 IU/mL; therefore, test serum samples with concentrations of 0.5–1.5 IU/mL will be read as false negatives (41). Jemima et al. (31) further developed the LAT, using the recombinant glycoprotein of RABV expressed in an insect cell system as the antigen and decreasing the detection limit from 2 to 0.88 IU/mL.

In addition to using traditional latex beads, Wang et al. (71) developed a magnetic protein microbead-aided indirect fluoroimmunoassay. The principle of this assay is based on (i) preparing magnetic BSA/γ-Fe2O3 microbeads carrying avidin; (ii) biotinylated virus antigen grafting onto the beads (antigen conjugate with avidin); (iii) antibody in the test serum binding to the antigen on the beads to form an antigen–antibody complex; (iv) FITC-labeled detection antibody conjugating with the antigen–antibody complex; and (v) the level of antibody being reflected by the system fluorescence intensities. This method not only provides a convenient means for rapid and sensitive assays but also utilizes the advantages of excellent biodegradability, nontoxicity, and biocompatibility of magnetic protein microbeads.

Electrochemiluminescence assay

An immunosensor is a type of affinity solid-state-based biosensor. In an immunosensor, the antigen and antibody form a stable antigen and antibody immunocomplex, generating a measurable signal given by a transducer (23,45,47). Immunosensors have high sensitivity and selectivity (15,72) and are widely used for detecting specific biomarkers (18,36,88). The main goal of using an immunosensor to detect special antibodies or antigens is to build a sensitive and reliable detection platform.

Ma et al. (40) developed an ELISA-based electrochemiluminescence (ECL) assay to detect RABV antibodies. The ECL assay is based on Meso Scale Discovery technology, which uses ECL labels and carbon electrode plates to detect antigen and antibody complexes. Compared with the conventional ELISA, the ECL assay demonstrates higher sensitivity. Although the ECL method produces a faster and more sensitive result, it has the inherent limitations of all ELISA methods versus a functional test.

Other methods

Other methods are also used to detect RABV antibodies (59).

The simplified fluorescence inhibition microtest (19) is easy to perform and more readable. This method has been used to analyze 22,000 serum samples each year by three Brazilian laboratories. It can also be used to detect low concentrations of RVNAs in wild carnivores and other wildlife in Brazil (6,32,54).

The immunoperoxidase inhibition assay (IIA) (2) is another alternative test for detecting RABV antibodies in serum samples. It is a blocking assay based on the reaction between test serum samples and RABV antigens in infected cells. The test is low in cost and does not require specific equipment, and the results are easily readable. IIA may be advantageous over other similar methods designed to detect RABV-specific binding antibodies because it can be easily introduced into most laboratories, provided basic cell culture facilities are available.

Counter immunoelectrophoresis (17) is another technique for titration of RABV antibodies. It detects antibodies of the IgG class, viral antiglycoproteins, the main immunoglobulins involved in the response against the virus. It can determine the neutralizing potential of analyzed serum samples (16).

Comparison of the Different Detection Methods of RABV Antibodies

A literature search of PubMed (www.ncbi.nlm.nih.gov/pubmed) and Embase (www.embase.com) databases was conducted for comparison of different detection methods of RABV antibodies. The following key words were used: “rabies antibody,” “RFFIT,” “FAVN,” and “ELISA.” Studies were included in this review if (i) the reference technique was RFFIT or FAVN, (ii) the samples were serum or body fluid samples from humans or animals, (iii) the sensitivity and specificity of the method were provided or the raw data were sufficient to calculate the sensitivity and specificity, and (iv) the study was published in English. Studies were excluded if they were duplicate literature.

Sensitivity was calculated by the following formula: (true positive/[true positive + false negative]). Specificity was calculated using the following formula: (true negative/[false positive + true negative]). The 95% confidence intervals were calculated using the Wilson/Brown method. Pearson's product–moment correlation and Bland–Altman agreement analysis were used to evaluate the performance of tests with continuous scales. Cohen's kappa and McNemar's test were used to evaluate the performance of tests with nominal scales. Pearson's product–moment correlation was used to examine the correlation between the new method and the gold standard. Bland–Altman analysis and Cohen's kappa were used to evaluate agreement between them. The kappa values were interpreted as follows: <0.00, poor agreement; 0.00–0.20, slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1.00, almost perfect agreement (37,55). McNemar's test was used to evaluate differences, and p-values ≤0.05 were considered statistically significant. Statistical analysis was performed using GraphPad 8.0 and SPSS 22.0.

Result

Twenty studies were included in this review. The performance of RABV antibody detection methods in human and animal samples is shown in Tables 1 and 2. Of 34 methods, 5 demonstrated unsatisfactory sensitivity or specificity. The others exhibited sensitivity and specificity of ≥75%. The correlation coefficient for five of eight methods was >0.8. The Bland–Altman mean bias of five of five methods was <±2.0. The kappa values of 25 of 28 methods were higher than 0.4, demonstrating at least moderate agreement.

Table. 1. The Performance of Methods Detecting Rabies Virus Antibodies in Human Samples

Study and year	Detection technique	Reference technique	Sample size	Sensitivity	Specificity	Correlation coefficient	Bland–Altman (mean bias)	Kappa value	McNemar test (p value)	Cutoff value
Burgado et al. (5)	HTNT (virus neutralization assay)	RFFIT	135	100.0% (0.9408–1.000)	100.0% (0.9507–1.000)	r = 0.88	1.17^a	/	/	0.5 IU/mL
Khawplod et al. (33)	GFP (virus neutralization assay)	RFFIT	25	100.0% (0.7961–1.000)	80.0% (0.4902–0.9645)	r = 0.988	−0.28^a	/	/	0.5 IU/mL
Ma et al. (40)	ECL (binding assay)	RFFIT	120	100.0% (0.9434–1.000)	94.6% (0.8539–0.9854)	/	/	kappa = 0.950	0.250	0.15 IU/mL
Zhao et al. (87)	ELISA (binding assay) Indirect based	RFFIT	415	94.2% (0.9144–0.9610)	78.9% (0.5667–0.9149)	/	/	kappa = 0.496	< 0.001	0.5 IU/mL
Zhao et al. (87)	ELISA (binding assay) Indirect based	RFFIT	415	91.1% (0.8770–0.9369)	100.0% (0.9442–1.000)	/	/	kappa = 0.763	< 0.001	1.5 IU/mL
Shiota et al. (62)	RAPINA (binding assay)	RFFIT	115	88.7% (0.7742–0.9471)	91.9% (0.8247–0.9651)	/	/	kappa = 0.807	> 0.999	0.5 IU/mL
Nishizono et al. (51)	RAPINA (qualitative evaluation) (binding assay)	RFFIT	323	98.9% (0.9612–0.9981)	99.3% (0.9604–0.9996)	/	/	kappa = 0.981	> 0.999	0.5 IU/mL
Nishizono et al. (51)	RAPINA (semiquantitative evaluation) (binding assay)	RFFIT	80	98.1% (0.9023–0.9991)	100.0% (0.8713–1.000)	/	/	kappa = 0.972	> 0.999	0.5 IU/mL
Nishizono et al. (51)	RAPINA (semiquantitative evaluation) (binding assay)	RFFIT	80	88.5% (0.7102–0.9600)	90.7% (0.8009–0.9598)	/	/	kappa = 0.777	0.727	4.0 IU/mL
Wang et al. (70)	ICTS (binding assay)	FAVN	756	90.4% (0.8699–0.9304)	98.2% (0.9634–0.991)	/	/	kappa = 0.889	< 0.001	0.5 IU/mL
Welch et al. (78)	ELISA (binding assay) Indirect based	RFFIT	94	94.1% (0.8091–0.9895)	95.8% (0.8602–0.9926)	0.685	/	kappa = 0.900	> 0.999	0.5 EU/mL
Welch et al. (78)	ELISA (binding assay) Focus diagnostics in-house-developed rabies Antibody detection by ELISA	RFFIT	94	91.7% (0.7817–0.9713)	73.0% (0.5702–0.8460)	0.685	/	kappa = 0.645	0.092	0.5 IU/mL
Feyssaguet et al. (20)	ELISA (binding assay) Indirect based	RFFIT	655	98.9% (0.9630–0.9982)	99.4% (0.9811–0.9982)	/	/	kappa = 0.982	> 0.999	0.5 EU/mL
Feyssaguet et al. (20)	ELISA (binding assay) Indirect based	RFFIT	655	84.4% (0.7357–0.9129)	96.9% (0.9524–0.9806)	/	/	kappa = 0.770	0.185	4.0 EU/mL

Titers were log transformed.

ECL, electrochemiluminescence assay; ELISA, enzyme-linked immunosorbent assay; FAVN, fluorescent antibody virus neutralization; GFP, green fluorescent protein; GMT, geometric mean titer; HTNT, high throughput neutralization test; ICTS, immunochromatographic test strip; RAPINA, rapid neutralizing antibody test; RFFIT, rapid fluorescent focus inhibition test.

Table. 2. The Performance of Methods Detecting Rabies Virus Antibodies in Animal Samples

Study and year	Detection technique	Reference technique	Sample size	Sensitivity	Specificity	Correlation coefficient	Bland–Altman (mean bias)	Kappa value	McNemar Test (p value)	Cutoff value
Burgado et al. (5)	HTNT (virus neutralization assay)	RFFIT	42	100.0% (0.8928–1.000)	100.0% (0.7225–1.000)	r = 0.96	0.29^a	/	/	0.5 IU/mL
Khawplod et al. (33)	GFP (virus neutralization assay)	RFFIT	18	90.0% (0.5958–0.9949)	100.0% (0.6756–1.000)	r = 0.948	0.13^a	/	/	0.5 IU/mL
Qin et al. (57)	GFP (virus neutralization assay)	RFFIT	27	100.0% (0.8318–1.000)	100% (0.6756–1.000)	r = 0.989	/	/	/	0.5 IU/mL
Smith et al. (65)	Microneutralization test (virus neutralization assay)	RFFIT	129	96.0% (0.8654–0.9929)	100.0% (0.9536–1.000)	r = 0.74	-1.4^b	/	/	0.1 IU/mL
Ma et al. (40)	ECL (binding assay)	RFFIT	180	100.0% (0.9487–1.000)	95.4% (0.8971–0.9802)	/	/	kappa = 0.943	0.063	0.15 IU/mL
Ma et al. (40)	ECL (binding assay)	RFFIT	167	96.2% (0.8111–0.9980)	95.7% (0.9103–0.9804)	/	/	kappa = 0.852	0.125	0.15 IU/mL
Nishizono et al. (51)	RAPINA (qualitative evaluation) (binding assay)	RFFIT	527	97.7% (0.9558–0.9884)	99.4% (0.9683–0.9997)	/	/	kappa = 0.962	0.039	0.5 IU/mL
Nishizono et al. (51)	RAPINA (semiquantitative evaluation) (binding assay)	RFFIT	153	96.3% (0.9094–0.9856)	100.0% (0.9197–1.000)	/	/	kappa = 0.938	0.125	0.5 IU/mL
Nishizono et al. (51)	RAPINA (semiquantitative evaluation) (binding assay)	RFFIT	153	75.0% (0.6122–0.8508)	97.1% (0.9193–0.9922)	/	/	kappa = 0.760	0.035	4.0 IU/mL
Jemima et al. (31)	LAT (binding assay)	RFFIT	228	96.0% (0.9160–0.9817)	100.0% (0.9525–1.000)	/	/	kappa = 0.942	0.031	0.88 IU/mL
Cliquet et al. (11)	ELISA (binding assay) Indirect based	FAVN	556	86.4% (0.8156–0.9019)	96.9% (0.9401–0.9848)	/	/	kappa = 0.837	< 0.001	0.6 EU/mL
Wasniewski et al. (76)	ELISA (binding assay) Indirect based	FAVN	593	78.2% (0.7454–0.8153)	100.0% (0.9398–1.000)	/	/	kappa = 0.421	< 0.001	0.5 EU/mL
Wasniewski et al. (76)	ELISA (binding assay) Indirect based	FAVN	593	86.9% (0.8373–0.8947)	96.7% (0.8864–0.9941)	/	/	kappa = 0.556	< 0.001	0.3 EU/mL
Servat et al. (61)	ELISA (binding assay) Indirect based	FAVN	588	88.9% (0.8579–0.9135)	98.9% (0.9422–0.9995)	/	/	kappa = 0.708	< 0.001	0.5 EU/mL
De Benedictis et al. (13)	ELISA (binding assay) Indirect based	FAVN	705	53.5% (0.4895–0.5800)	74.9% (0.6860–0.8048)	/	/	kappa = 0.23	/	0.5 EU/mL
De Benedictis et al. (13)	ELISA (binding assay) Competition based	FAVN	682	83.1% (0.7944–0.8638)	25.9% (0.2014–0.3247)	/	/	kappa = 0.10	/	40%
De Benedictis et al. (13)	ELISA (binding assay) Competition based	FAVN	682	56.9% (0.5237–0.6147)	62.9% (0.5603–0.6955)	/	/	kappa = 0.17	/	70%
Wasniewski et al. (75)	ELISA (binding assay) Competition based	FAVN	628	98.4% (0.9689–0.9916)	67.6% (0.5557–0.7772)	/	/	kappa = 0.719	0.024	40%
Wasniewski et al. (74)	ELISA (binding assay) Competition based	FAVN	1,016	84.7% (0.8151–0.8735)	99.5% (0.9799–0.9991)	/	/	kappa = 0.808	< 0.001	40%
Yang et al. (83)	ELISA (binding assay) Double-antigen sandwich based	FAVN	500	92.7% (0.8896–0.9521)	95.2% (0.9153–0.9727)	/	/	kappa = 0.875	0.150	0.9 EU/mL

Titers were log transformed.

Titers were GMT transformed.

LAT, latex agglutination test.

Discussion

Detection of RABV antibodies is a good indicator of a person's degree of immunity during the antirabies treatment or after vaccination. Evidence has indicated that virus neutralization assays are more suitable for laboratory quantitative detection and binding assays are more suitable for fast testing. In this article, we reviewed RABV antibody-detecting technologies and analyzed their performance.

The results of the present review demonstrated that most of the methods had sensitivity and specificity of ≥75%, indicating acceptable results. In fact, the cutoff value is an important influence of method performance. For improving the diagnostic power of a test, sensitivity and specificity could be altered by changing the cutoff levels (29). A similar conclusion could be confirmed in this review [the results of (87) and (51)]. For some commercial ELISA kits, different kit batches, although from the same manufacturer, have different cutoff values. When the experimenters use the new method or new kits to test samples, the cutoff value should be chosen objectively and carefully to obtain satisfactory test results. In general, higher sensitivity and specificity are better. Because most diagnostic tests have high specificity and low sensitivity or low specificity and high sensitivity, their value is mainly evaluated according to the application purpose of diagnostic tests. If both sensitivity and specificity are considered, receiver operating characteristic curve analysis can be used to improve diagnostic performance.

Most of the methods included in this review (5,33,57) used Pearson's product–moment correlation test and the r-value to determine correlation. Fewer studies (66) used paired t-tests to determine correlation. However, using only paired t-tests to analyze correlation is not comprehensive. Even if the p-value is >0.05, this may not mean that the results of the two methods are consistent; there is a possibility of type II error. Meanwhile, the p-value of a paired t-test is qualitative. We still must use other statistical methods to quantify the agreement between the test method and the gold standard. Hence, we performed Bland–Altman analyses on quantitative outcomes. Bland–Altman analysis is an up-to-date statistical technique checking agreement between two quantitative measurements by constructing limits of agreement. These statistical limits are calculated by using the mean and standard deviation of the differences between two measurements. Bland–Altman analysis quantifies the bias and a range of agreement. However, in this review, we calculated only the bias due to the lack of raw data in some studies. Agreement of qualitative outcomes was analyzed using McNemar's test and Cohen's kappa. In normal cases, if the p-value of McNemar's test was insignificant (≥0.05) and p-value of the kappa statistic was significant (<0.05), this was considered as being in agreement, and the kappa value determined the strength of agreement. However, in some analyses, while the kappa value suggested that the two methods had relatively high levels of agreement, the significant generalized McNemar's test provided additional evidence of disagreement (p < 0.05) between them. The possible reason for this phenomenon is that McNemar's test was only used to calculate the difference of two methods; the agreement and the sample size were not considered. If the total sample (n), true-positive sample (a), and true-negative sample (d) sizes are large and the false-positive sample (b) and false-negative sample (c) sizes are small, McNemar's test loses its practical application. Therefore, in this phenomenon, the kappa value should be chosen first to evaluate the performance of the new method. According to the results of the human sample group in this review (Table 1), the p-values for 9 of 12 methods were >0.05 and the kappa values were >0.6. This indicates good performance of the methods for human samples. In addition, for the group of animal samples, although the p-values for only 4 of 16 methods were >0.05, the kappa values for 13 of 16 methods were >0.4. According to the interpretation of kappa value, a kappa value >0.4 indicated at least moderate agreement of two methods. Therefore, most of the results for animal samples are acceptable.

Indirect ELISA, sandwich ELISA, and competitive ELISA were used to detect the RABV antibody. In this review, the performance of 10 indirect ELISAs, 4 competitive ELISAs, and 1 sandwich ELISA was evaluated (Tables 1 and 2). Our results show that of 10 indirect ELISAs, 9 have acceptable results (sensitivity: 78.2–98.96%, specificity: 78.9–100%). One sandwich ELISA has a satisfactory result (sensitivity: 92.67%, specificity: 95.15%). The results of two (13) of four competitive ELISAs are unsatisfactory. The authors demonstrated that a nonhomogeneous or insufficient concentration of antigens coating the ELISA microplates or the quality of the sample could partially explain the unsatisfactory results.

Sample quality is crucial for detecting rabies antibodies. Poor-quality serum samples, such as field animal serum samples, may produce unsatisfactory results in virus neutralization assays (5,33,40,51). The RFFIT and FAVN test are too sensitive to cytotoxicity (12). Shiraishi et al. (63) reported that some factors, such as some anticoagulants, anti-BHK-21 cell IgG antibodies, and serum storage at temperature >25°C, could induce cytotoxicity. Cytotoxicity has less influence in the microneutralization test than in conventional RFFIT (65). The first dilution of a microneutralization test is 1:10 (higher than in conventional RFFIT); therefore, the microneutralization test is less sensitive to cytotoxicity. Bedekoviće et al. (3) developed a modified FAVN test to eliminate this cytotoxic effect. ELISAs are usually less susceptible to interfering substances because the serum dilution used in the assay is generally higher (46). ELISAs coupled with filter paper have shown satisfactory results for assessing poor-quality field animal serum samples (12).

As reviewed in the present study, developments in virus neutralization assays include inserting GFP into the virus genome to recombine the challenge virus. In comparison with conventional RFFIT, which requires laboratory operators to subjectively read the percentages of the fluorescent foci in the wells, this recombined challenge virus, combined with array scanning or ultraviolet microscopy with a cell imaging system, makes the results more precise and accurate. Binding assays, such as ELISA, RAPINA, and LAT, detect all RABV-specific antibodies or focus on neutralizing antibodies (the c-ELISA specifically evaluates the RABV-neutralizing potency of serum) (34) with good performance. By purifying and expressing high-quality and high-concentration RABV glycoproteins, these methods yield results consistent with those of virus neutralization assays. ECL and other immunosensors reflect a potentially valuable testing method for detection of RABV antibodies. Based on the characteristics of biomaterials, a detection platform with high specificity and sensitivity could be constructed. Such a platform is very sensitive to the signals generated by binding events.

In conclusion, any new technology should be considered carefully and objectively before being used as an appropriate and applicable alternative. More research should be performed to verify the appropriateness and applicability of methods tested for detecting RABV antibodies. RABV antibody detection technologies still have practical value and promising prospects.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

Alhajj

, and Farhana

Enzyme linked immunosorbent assay. In: StatPearls. Treasure Island (FL): StatPearls Publishing 2021.

Batista

, Lima

, Maletich

, et al. Immunoperoxidase inhibition assay for rabies antibody detection. J Virol Methods, 2011; 174:65–68.

Bedeković

, Lemo

, Lojkić

, et al. Modification of the fluorescent antibody virus neutralisation test—elimination of the cytotoxic effect for the detection of rabies virus neutralising antibodies. J Virol Methods, 2013; 189:204–208.

Bhatt

, Hattwick

MAW

, Gerdsen

, et al. Human rabies: diagnosis, complications, and management. Am J Dis Child, 1974; 127:862–869.

Burgado

, Greenberg

, Niezgoda

, et al. A high throughput neutralization test based on GFP expression by recombinant rabies virus. PLoS Negl Trop Dis, 2018; 12:e0007011.

Casagrande

, Favaro

, Carvalho

, et al. Rabies surveillance in bats in Northwestern State of São Paulo. Rev Soc Bras Med Trop, 2014; 47:709–715.

Centers for Disease Control and Prevention Human rabies prevention United

States

, 2008. Recommendations of the advisory committee on immunization practices. MMWR, 2008; 57:1–36.

Centers for Disease Control and Prevention

Recovery of a patient from clinical rabies—California

, 2011. MMWR 2012;61:61–65.

Chosewood

, Wilson DE; Centers for Disease Control and Prevention (U.S.); National Institutes of Health (U.S.). Biosafety in Microbiological and Biomedical Laboratories. 5th ed. U.S. Dept. of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention For sale by the Supt. of Docs. Washington, DC: U.S. G.P.O., 2009.

10.

Cliquet

, Aubert

. and Sagne L. Development of a fluorescent antibody virus neutralisation test (FAVN test) for the quantitation of rabies-neutralising antibody. J Immunol Methods, 1998; 212:79–87.

11.

Cliquet

, McElhinney

, Servat

, et al. Development of a qualitative indirect ELISA for the measurement of rabies virus-specific antibodies from vaccinated dogs and cats. J Virol Methods, 2004; 117:1–8.

12.

Cliquet

, Sagné

, Schereffer

, et al. ELISA test for rabies antibody titration in orally vaccinated foxes sampled in the fields. Vaccine, 2000; 18:3272–3279.

13.

De Benedictis

, Mancin

, Cattoli

, et al. Serological methods used for rabies post vaccination surveys: an analysis. Vaccine, 2012; 30:5611–5615.

14.

Debnath

, Pathak

, Ramamurthy

, et al. Serological profiling of rabies antibodies by enzyme-linked immunosorbent assay and its comparative analysis with rapid fluorescent focus inhibition test in mouse model. Vet World, 2019; 12:126–130.

15.

Diaconu

, Cristea

, Hârceagă

, et al. Electrochemical immunosensors in breast and ovarian cancer. Clin Chim Acta, 2013; 425:128–138.

16.

Díaz

, and Myers

. Comparison between a modified counterimmunoelectrophoresis test and the indirect fluorescent-antibody test for detection of antibodies to rabies virus in human sera. J Clin Microbiol, 1981; 14:446–448.

17.

Díaz

, and Myers

. Determination of serum neutralization antibodies to rabies virus by a modified counterimmunoelectrophoresis test. J Clin Microbiol, 1980; 12:175–179.

18.

Farzin

, Shamsipur

, Samandari

, et al. HIV biosensors for early diagnosis of infection: the intertwine of nanotechnology with sensing strategies. Talanta, 2020; 206:120201.

19.

Favoretto

, Carrieri

, Tino

, Zanetti

, and Pereira

. Simplified fluorescent inhibition microtest for the titration of rabies neutralizing antibodies. Rev Inst Med Trop Sao Paulo, 1993; 35:171–175.

20.

Feyssaguet

, Dacheux

, Audry

, et al. Multicenter comparative study of a new ELISA, PLATELIA RABIES II, for the detection and titration of anti-rabies glycoprotein antibodies and comparison with the rapid fluorescent focus inhibition test (RFFIT) on human samples from vaccinated and non-vaccinated people. Vaccine, 2007; 25:2244–2251.

21.

Fisher

, Streicker

, and Schnell

. The spread and evolution of rabies virus: conquering new frontiers. Nat Rev Microbiol, 2018; 16:241–255.

22.

Fitzgerald

EA.

Potency test of antirabies serum and immunoglobulin. In: Meslin FX, Kaplan MM, and Koprowsky H, ed. Laboratory Techniques in Rabies. 4th ed. Geneva: World Health Organization, 1996:417—22.

23.

Ghindilis

, Atanasov

, Wilkins

, and Wilkins

. Immunosensors: electrochemical sensing and other engineering approaches. Biosens Bioelectron, 1998; 13:113–131.

24.

Gilbert

, Petersen

, et al. Evidence of rabies virus exposure among humans in the Peruvian Amazon. Am J Trop Med Hyg, 2012; 87:206–215.

25.

Gispen

, and Saathof

. Neutralizing and fluorescent antibody response in man after anti rabies treatment with suckling rabbit brain vaccine. Archiv fur die GesamteVirusforschung, 1965; 15:377–386.

26.

Gold

, Donnelly

, Nouvellet

, and Woodroffe

. Rabies virus-neutralising antibodies in healthy, unvaccinated individuals: what do they mean for rabies epidemiology?. PLoS Negl Trop Dis, 2020; 14:e0007933.

27.

Goldwasser

, and Kissling

. Fluorescent antibody staining of street and fixed rabies virus antigens. Proc Soc Exp Biol Med, 1958; 98:219–223.

28.

Hanlon

, Niezgoda

, and Rupprecht

. Postexposure prophylaxis for prevention of rabies in dogs. Am J Vet Res, 2002; 63:1096–1100.

29.

Holle

, Hellmich

, Backes

, et al. Variations in performance characteristics of commercial enzyme immunoassay kits for detection of antineutrophil cytoplasmic antibodies: what is the optimal cut-off?. Ann Rheum Dis, 2005; 64:1773–1779.

30.

Hostnik

The modification of fluorescent antibody virus neutralization (FAVN) test for the detection of antibodies to rabies virus. J Vet Med B Infect Dis Vet Public Health, 2000; 47:423–427.

31.

Jemima

, Manoharan

, and Kumanan

. Development and evaluation of a recombinant-glycoprotein-based latex agglutination test for rabies virus antibody assessment. Arch Virol, 2014; 159:1987–1993.

32.

Jorge

, Pereira

, Morato

, et al. Detection of rabies virus antibodies in Brazilian free-ranging wild carnivores. J Wildl Dis, 2010; 46:1310–1315.

33.

Khawplod

, Inoue

, Shoji

, et al. A novel rapid fluorescent focus inhibition test for rabies virus using a recombinant rabies virus visualizing a green fluorescent protein. J Virol Methods, 2005; 125:35–40.

34.

Korimbocus

, Dehay

, Tordo

, et al. Development and validation of a quantitative competitive ELISA for potency testing of equine anti rabies sera with other potential use. Vaccine, 2016; 34:3310–3316.

35.

Kuzmin

, Niezgoda

, Franka

, et al. Lagos bat virus in Kenya. J Clin Microbiol, 2008; 46:1451–1461.

36.

Kveton

, Blsakova

, Lorencova

, et al. A graphene-based glycan biosensor for electrochemical label-free detection of a tumor-associated antibody. Sensors (Basel), 2019; 19:5409.

37.

Landis

, and Koch

. The measurement of observer agreement for categorical data. Biometrics, 1977; 33:159–174.

38.

Leffingwell

, and Irons

. Rabies antibodies in human serums titrated by the indirect FA method. Publ Health Rep, 1965; 80:999–1004.

39.

Louie

, Dobkin

, Meyer

, et al. Measurement of rabies antibody: comparison of the mouse neutralization test (MNT) with the rapid fluorescent focus inhibition test (RFFIT). J Biol Stand, 1975; 3:365–373.

40.

, Niezgoda

, Blanton

, et al. Evaluation of a new serological technique for detecting rabies virus antibodies following vaccination. Vaccine, 2012; 30:5358–5362.

41.

Madhusudana

, and Saraswati

. Development and evaluation of a latex agglutination test for rabies antibodies. J Clin Virol, 2003; 27:129–135.

42.

Messenger

, Rupprecht

. Chapter fourteen—an indirect fluorescent antibody test for the serological detection of rabies virus immunoglobulin G and immunoglobulin M antibodies. Curr Lab Tech Rabies Diagn Res Prev, 2015:2:157–173.

43.

Meza

, Broos

, Becker

, et al. Predicting the presence and titer of rabies virus neutralizing antibodies from low-volume serum samples in low-containment facilities. bioRxiv 2020.04.24.060095.

44.

Moeschler

, Locher

, Conzelmann

, Krämer

, and Zimmer

. Quantification of lyssavirus-neutralizing antibodies using vesicular stomatitis virus pseudotype particles. Viruses, 2016; 8:254.

45.

Mollarasouli

, Kurbanoglu

, and Ozkan

. The role of electrochemical immunosensors in clinical analyzed. Biosensors (Basel), 2019; 9:86.

46.

Moore

, and Hanlon

. Rabies-specific antibodies: measuring surrogates of protection against a fatal disease. PLoS Negl Trop Dis, 2010; 4:e595.

47.

Morgan

, Newman

, and Price

. Immunosensors: technology and opportunities in laboratory medicine. Clin Chem, 1996; 42:193–209.

48.

Muhamuda

, Madhusudana

, and Ravi

. Development and evaluation of a competitive ELISA for estimation of rabies neutralizing antibodies after post-exposure rabies vaccination in humans. Int J Infect Dis, 2007; 11:441–445.

49.

Nicholson

, and Prestage

. Enzyme-linked immunosorbent assay: a rapid reproducible test for the measurement of rabies antibody. J Med Virol, 1982; 9:43–49.

50.

Nie

, Wu

, Ma

, et al. Development of in vitro and in vivo rabies virus neutralization assays based on a high-titer pseudovirus system. Sci Rep, 2017; 7:42769.

51.

Nishizono

, Yamada

, Khawplod

, et al. Evaluation of an improved rapid neutralizing antibody detection test (RAPINA) for qualitative and semiquantitative detection of rabies neutralizing antibody in humans and dogs. Vaccine, 2012; 30:3891–3896.

52.

OIE: Rabies (infection with rabies virus). In: Manual of Diagnostic Tests and Vaccines for Terrestrial Animals. 7th ed. Paris: World Organisation for Animal Health, 2012:1–28.

53.

Omodo

, Ar Gouilh

, Mwiine

, et al. Rabies in Uganda: rabies knowledge, attitude and practice and molecular characterization of circulating virus strains. BMC Infect Dis, 2020; 20:200.

54.

Onuma

SSM

, Chaves

, Souza

MCC

, et al. Serological and molecular investigation of viral agents in free-living jaguars in the Pantanal of Mato Grosso, Brazil. Braz J Vet Res An Sci, 2016; 53:270–279.

55.

Pålsson

, Kostogiannis

, and Ageberg

. Combining results from hip impingement and range of motion tests can increase diagnostic accuracy in patients with FAI syndrome. Knee Surg Sports Traumatol Arthrosc, 2020; 28:3382–3392.

56.

Preiss

, Chanthavanich

, Chen

, et al. Post-exposure prophylaxis (PEP) for rabies with purified chick embryo cell vaccine: a systematic literature review and meta-analyzed. Expert Rev Vaccines, 2018; 17:525–545.

57.

Qin

, Volokhov

, Rodionova

, et al. A new recombinant rabies virus expressing a green fluorescent protein: a novel and fast approach to quantify virus neutralizing antibodies. Biologicals, 2019; 59:56–61.

58.

Realegeno

, Niezgoda

, Yager

, et al. An ELISA-based method for detection of rabies virus nucleoprotein-specific antibodies in human antemortem samples. PLoS One, 2018; 13:e0207009.

59.

Rupprecht

, Fooks

, and Abela-Ridder

. Laboratory Techniques in Rabies. 5th ed. Geneva, Switzerland: World Health Organization, 2018.

60.

Servat

, Cliquet

; and OIE Reference Laboratory for Rabies; WHO. Collaborative study to evaluate a new ELISA test to monitor the effectiveness of rabies vaccination in domestic carnivores. Virus Res, 2006; 120:17–27.

61.

Servat

, Feyssaguet

, Blanchard

, et al. A quantitative indirect ELISA to monitor the effectiveness of rabies vaccination in domestic and wild carnivores. J Immunol Methods, 2007; 318:1–10.

62.

Shiota

, Mannen

, Matsumoto

, et al. Development and evaluation of a rapid neutralizing antibody test for rabies. J Virol Methods, 2009; 161:58–62.

63.

Shiraishi

, Nishimura

, Nakashima

, et al. Determination of critical factors causing cytotoxicity in the virus neutralization test. J Virol Methods, 2014; 199:46–52.

64.

Smith

, Yager

, and Baer

. A rapid reproducible test for determining rabies neutralizing antibody. Bull World Health Organ, 1973; 48:535–541.

65.

Smith

, and Gilbert

. Comparison of a micro-neutralization test with the rapid fluorescent focus inhibition test for measuring rabies virus neutralizing antibodies. Trop Med Infect Dis, 2017; 2:24.

66.

Tang

, Lu

, Wei

, et al. A recombinant rabies virus expressing a phosphoprotein-eGFP fusion is rescued and applied to the rapid virus neutralization antibody assay. J Virol Methods, 2015; 219:75–83.

67.

Thomas

, Sikes

, Ricker

. Evaluation of indirect fluorescent antibody technique for detection of rabies antibody in human sera. J Immunol, 1963; 91:721–723.

68.

Tian

, Chen

, and Yan

. Clinical features of rabies patients with abnormal sexual behaviors as the presenting manifestations: a case report and literature review. BMC Infect Dis, 2019; 19:679.

69.

Timiryasova

, Luo

, Zheng

, et al. Rapid fluorescent focus inhibition test optimization and validation: improved detection of neutralizing antibodies to rabies virus. J Immunol Methods, 2019; 474:112626.

70.

Wang

, Feng

, Yang

, et al. A rapid immunochromatographic test strip for detecting rabies virus antibody. J Virol Methods, 2010; 170:80–85.

71.

Wang

, Ren

, Tu

, et al. Magnetic protein microbead-aided indirect fluoroimmunoassay for the determination of canine virus specific antibodies. Biosens Bioelectron, 2011; 26:3353–3360.

72.

Wang

, Xu

, Zhang

, et al. Electrochemical sensors for clinic analyzed. Sensors (Basel), 2008; 8:2043–2081.

73.

Wasniewski

, Barrat

, Combes

, et al. Use of filter paper blood samples for rabies antibody detection in foxes and raccoon dogs. J Virol Methods, 2014; 204:11–16.

74.

Wasniewski

, Cliquet

. Evaluation of ELISA for detection of rabies antibodies in domestic carnivores. J Virol Methods, 2012; 179:166–175.

75.

Wasniewski

, Guiot

, Schereffer

, et al. Evaluation of an ELISA to detect rabies antibodies in orally vaccinated foxes and raccoon dogs sampled in the field. J Virol Methods, 2013; 187:264–270.

76.

Wasniewski

, Labbe

, Tribout

, et al. Evaluation of a rabies ELISA as an alternative method to seroneutralisation tests in the context of international trade of domestic carnivores. J Virol Methods, 2014; 195:211–220.

77.

Webster

, and Dawson

. Early diagnosis of rabies by mouse inoculation. Measurement of humoral immunity to rabies by mouse protection test. Proc Soc Exp Biol Med, 1935; 32:570–573.

78.

Welch

, Anderson

, and Litwin

. An evaluation of two commercially available ELISAs and one in-house reference laboratory ELISA for the determination of human anti-rabies virus antibodies. J Med Microbiol, 2009; 58(Pt 6):806–810.

79.

WHO Expert Committee on biological standardization. World Health Organ Tech Rep Ser 2007;941:1–340.

80.

Wijaya

, Tham

CYL

, Chan

YFZ

, et al. An accelerated rabies vaccine schedule based on toll-like receptor 3 (TLR3) agonist PIKA adjuvant augments rabies virus specific antibody and T cell response in healthy adult volunteers. Vaccine, 2017; 35:1175–1183.

81.

World Health Organization. Rabies vaccine: WHO position paper. Wkly Epidemiol Rec, 2010; 32:309–320.

82.

World Health

Organization

. WHO position paper on rabies vaccine. Grading of scientific evidence, 2010. https://www.who.int/publications/i/item/who-wer8532 (accessed September 7, 2021 ).

83.

Yang

, Zhao

, Hu

, et al. A novel double-antigen sandwich enzyme-linked immunosorbent assay for measurement of antibodies against rabies virus. Clin Vaccine Immunol, 2006; 13:966–968.

84.

Zhang

, Guo

, and Wang

. Immunochromatographic lateral flow strip tests. Methods Mol Biol, 2009; 504:169–183.

85.

Zhang

, Wang

, Yang

, et al. Development of an immunochromatographic lateral flow test strip for detection of beta-adrenergic agonist Clenbuterol residues. J Immunol Methods, 2006; 312:27–33.

86.

Zhang

, Zhang

, Liu

, et al. Comparison between FAVN and RF FIT for rabies neutralizing antibody detection of sera from animals and human. Int J Lab Med 2015,36:1161–1163.

87.

Zhao

, Yu

, Shan

, et al. Rabies virus glycoprotein serology ELISA for measurement of neutralizing antibodies in sera of vaccinated human subjects. Vaccine, 2019; 37:6060–6067.

88.

Zhao

, Cai

, Zhu

, et al. A novel fluorescent and electrochemical dual-responsive immunosensor for sensitive and reliable detection of biomarkers based on cation-exchange reaction. Anal Chim Acta, 2020; 1096:61–68.

89.

Zoonoses control: WHO consultation on intradermal application of human rabies vaccines. Wkly Epidemiol Rec, 1995; 70:336–338.