Nicholas Vafai a, Kevin Self b, Bret Sheffield b, Sally Hojvat c, Aubrey Kusi-Appiah a, Patrick Vaughan b, Elliot Cowan c, Abbas Vafai a
Received 25 August 2022, Revised 11 January 2023, Accepted 18 January 2023, Available online 21 January 2023, Version of Record 25 January 2023.
Varicella zoster virus (VZV) causes childhood chickenpox, becomes latent in sensory ganglia and reactivates years later to cause shingles (Zoster) and postherpetic neuralgia in the elderly and immunosuppressed individuals. Serologic IgG tests can be used to determine if a person has antibodies to VZV from past varicella infection or had received varicella or zoster (shingles) vaccination. Commercial enzyme-linked immunosorbent assays (ELISAs) are currently used for the detection of VZV IgG antibodies in patient serum samples. However, ELISA tests require collection and processing of blood samples in a CLIA laboratory to separate serum or plasma for further testing. In this paper, we describe the development and testing of an antibody based Lateral Flow Immunochromatographic assay (LFA) device for the detection of VZV IgG in fingerstick whole blood. Analytical and clinical analyses were performed to compare the performance characteristics of the Viro VZV IgG LFA (VZV LFA) and the Diamedix VZV IgG ELISA. Analytical studies demonstrated the higher sensitivity of the VZV LFA compared to the ELISA by testing dilutions of the WHO VZV IgG serum International Standard. Clinical performance characteristics of the VZV LFA fingerstick whole blood assay were assessed at three point of care (POC) facilities by untrained users testing samples from 300 prospectively enrolled study subjects. VZV LFA results were compared with results obtained by testing serum samples obtained from the same study participants by the Diamedix VZV IgG ELISA. Two specimens with invalid results by the LFA assay were not included in the LFA performance calculations and nine equivocal ELISA results were included as positive for IgG results. The results from all three POC clinical sites demonstrated the higher sensitivity/positive percent agreement (PPA) (99.26%, 95% CI: 97.34–99.80) of the VZV LFA compared to the Diamedix VZV IgG ELISA (94.08%, 95% CI: 90.72–96.27). The specificity/negative percent agreement (NPA) of the VZV LFA compared to the ELISA test was calculated initially to be 39.29% (95% CI: 23.57–57.59) with 19 discordant test results out of 298 test results between the two assays (17 LFA positive/ELISA negative and two LFA negative/ELISA positive). The PPA and true NPA of the VZV LFA were determined by testing all 298 samples, including the discordant (19) and all concordant negative and positive (279) study subject serum samples, before and after blocking VZV gE antibody sites in the samples by spiking with VZV LFA gE capture antigen. The NPA improved to 100% (95% CI: 74.12–100) after the procedure when compared to the ELISA test results. The comparator ELISA PPA based on the spiking/blocking study remained as 94.08%, (95% CI: 90.72–96.27), comparable to test results from untreated samples.
The VZV LFA has been demonstrated to be simple and sufficiently robust for use in CLIA-waived POC facilities by untrained healthcare professionals and to detect VZV IgG in 20 min from fingerstick whole blood. The VZV LFA therefore provides a fast, reliable, and highly sensitive method of determining prior VZV viral infection or varicella and zoster vaccination status.
VZV, a member of the human herpesvirus family, causes two distinct clinical manifestations: childhood chickenpox (varicella) and shingles (zoster). Varicella is the outcome of the primary infection with VZV, whereas zoster is the result of VZV reactivation from latently infected sensory ganglia that occurs predominantly in aging and immunosuppressed individuals. Both varicella and zoster are more common and more severe in patients with underlying malignancies, steroid use, transplant patients and other immunocompromised individuals, and patients undergoing cytotoxic therapy or radiation therapy (Gelb, 1993; Weller, 1991; WelIer, 1979).
VZV is acquired through contact with persons with active chickenpox or shingles infection. Prior to the adoption of the first varicella vaccine approved by the FDA in 1995, there were an estimated 4 million cases of varicella each year in the United States (Gelb, 1993; Gelb, 1990). The incidence of varicella before the introduction of the varicella vaccine was higher in the African American population than in Americans of other ethnicities (Schmader, K. et al. 1995). Currently, varicella vaccine is administered to children along with a vaccine for measles, mumps and rubella. Over 90% of children in the US were immunized by 2010 with fewer than 100 cases per year of varicella being reported to the CDC (Gershon et al., 2021). The majority of the US adult population who have not been vaccinated by varicella vaccine either had natural childhood chickenpox or have been immunized by shingles vaccines (Gershon et al., 2021). The incidence of zoster in the US is estimated to be 1.2 million cases per year (Gelb, L.D. 1993). Studies reported from the United Kingdom, France, United States, Iceland, the Netherlands and Scotland all suggest that the incidence of zoster increases at age 50 and older (Steward and Human, 2007). VZV may reactivate to cause shingles due to various reasons including age, decline in cell-mediated immunity, and immunosuppression. In several studies among elderly, racially diverse populations in North Carolina, African Americans had a significantly lower lifetime occurrence and annual incidence of shingles than Americans of other ethnicities (Schmader et al., 1995).
Two attenuated VZV vaccines (VARIVAX®, ZOSTAVAX®) have been FDA approved to protect against chickenpox and shingles. An adjuvanted recombinant glycoprotein E (gE) subunit vaccine (SHINGRIX®) has also been approved to prevent shingles in individuals 50 years and older (Vafai, 2009; Vafai et al., 2000; Vafai, 1995; Vafai, 1994; Vafai, 1993).
Antibodies against VZV are currently detected using whole VZV antigens or mixed VZV glycoproteins by immunoassays such as enzyme-linked immunosorbent assays (ELISAs). A VZV glycoprotein E (gE)-based ELISA has also been used for the detection of VZV antibody in serum samples (Feyssaguet et al., 2020), however, glycoprotein-based antigens are not commercially available.
The purpose of this study was to describe the performance characteristics of a newly developed VZV IgG LFA. ELISA tests, including Diamedix ELISA, are reviewed for safety and effectiveness and cleared by the FDA for commercial use and are commonly used in the US clinical diagnostic laboratories to detect VZV antibodies in patients’ serum samples. In this study, the clinical trial the VZV LFA results were compared with the Diamedix ELISA. In addition, gE spiking was used to block gE antibody sites in the clinical study subjects’ serum samples to corroborate both the Diamedix ELISA positive/equivocal and the VZV LFA positive results and to demonstrate that the results from the clinical trials study were “true positive” or “true negative”.
Unlike the currently available ELISA test, the VZV LFA is a POC test for the qualitative detection of VZV IgG antibody in whole blood by a fingerstick at POC testing facilities. This assay can be used as an aid in the determination of a previous chickenpox or zoster infection, or successful chickenpox or shingles vaccination and the result is available immediately to a physician rather than waiting for an ELISA result from a CLIA certified laboratory.
VZV LFA is used for the detection of IgG antibodies against VZV by testing patient samples with a membrane-based lateral flow immunoassay test. The VZV LFA, unlike the available VZV ELISA tests, does not require collection and processing of venous whole blood samples but provides a simple and rapid POC test for the detection of VZV antibody in whole blood from a whole blood fingerstick.
The VZV LFA antibody detection test includes a truncated VZV glycoprotein E (gE) as the capture antigen. VZV gE is the most abundant and immunogenic VZV envelope glycoprotein and is capable of eliciting both humoral and cell-mediated immunity following VZV infection or vaccination against chickenpox or shingles. VZV gE is used in an adjuvanted subunit vaccine (SHINGRIX®) to prevent zoster or shingles in individuals aged 50 years and older. The VZV LFA test line consists of a highly purified truncated VZV recombinant gE protein.
2. Materials and methods
2.1. CHO gE expression and purification
The truncated VZV expression plasmids were prepared by subcloning Viro Research’s proprietary truncated VZV gene into a mammalian expression vector. The plasmid was transfected into the CHO-K1SP cells at 0.40 × 106 cells/mL for 48 h. The cells were then screened in the selective CD CHO medium with 25 μm MSX and 200× anti-clumping agent for another 48 h. The cell culture solution was centrifuged at 2000 rpm and the supernatant decanted for detection of truncated VZV. Detection of the truncated VZV was done using a dot blot protocol. Cultures which tested positive for the truncated VZV were used for protein production.
Truncated VZV production was done by culturing the selected CHO-K1SP cells for 48 h, centrifuging the cells at 2000 rpm, and decanting the supernatant for purification. A two-step protein purification was performed. First, affinity chromatography was performed using a Ni Sepharose Excel 60.4 mL column. Size exclusion chromatography was then performed using the Gel filtration column Superdex 26/600. The resulting protein solution obtained was sterilized by filtration through a 0.22 μm filter. The protein was analyzed for purity using SEC-HPLC.
Proteins were analyzed by SDS-PAGE and Western blot using standard protocols for molecular weight and purity measurements.
2.2. Specimen matrix comparison
This study used one (1) verification lot of VZV LFA to test plasma vs serum sample matrices and different sample anti-coagulants to determine the robustness of the VZV LFA serology. Whole blood samples containing K2 EDTA as well as serum samples from the same individual donor were used in this study. The blood samples with anti-coagulant were spun down and plasma were separated for testing. For each sample, 15 μL of sample was added to the sample port on the LFA cassette using the capillary tubes followed by three (3) drops of chase buffer. Devices were run for 20 min, then visual grades were read by two (2) different operators. Results were evaluated to determine variations in performance between donors, matrices and anti-coagulant.
2.3. Analytical sensitivity and limit of detection
The analytical performance of the VZV LFA was compared to the Diamedix ELISA test using a serial dilution series of the WHO VZV IgG International Standard, measured in IU/mL (WHO Guidelines, 2010). Serial dilutions (1× PBS) of the WHO standard were used to determine the sensitivity of VZV LFA in the detection of VZV IgG antibody. The dilutions were tested in 3 replicates for the LFA and 2 replicates in the ELISA. The comparator testing characterized specimens as negative or positive for VZV IgG and the VZV LFA also characterized them as positive or negative for VZV IgG. Results are shown in Table 2.
Potential cross-reactivity of VZV LFA was tested with the following seropositive serum samples (Discovery Life Sciences): HSV-1 (n = 10), HSV-2 (n = 10), cytomegalovirus (n = 8), rubella (n = 18), mumps (n = 8), rheumatoid factor (n = 2), Epstein-Barr Virus (n = 4), B burgdorferi (n = 6), T. gondii (n = 2), syphilis (n = 8), rubeola (n = 10), antinuclear antibody (n = 8), and chlamydia (n = 6). These samples were well characterized using FDA cleared assays by the vendor. To determine the effect of cross-reactivity, the well characterized samples were evaluated in the LFA and the ELISAs for each specific target. However, due to the prevalence of VZV antibodies in the United States population, the majority of cross-reactant serum samples were also reactive for VZV IgG when tested by the LFA. Therefore, all cross-reactant serum samples were pre-adsorbed with VZV gE, as described below, before testing to demonstrate that potentially cross-reacting analytes, at medically relevant concentrations present in a patient’s serum being tested for VZV IgG, will not cause a false VZV LFA test result with serum samples containing VZV IgG levels close to the test’s pre-determined cut-off.
2.4.1. Pre-adsorption with VZV gE
In order to block VZV gE antibody sites, cross-reactant serum samples were spiked with gE antigen to block VZV gE antibody sites. The concentration for the gE antigen spike was determined by taking known positive serum samples and spiking in different concentrations of gE antigen until the VZV LFA test line signal was competed away. Concentrations of the antigen spike were titrated from 0.1, 0.2, 0.3, 0.5, and 1 mg/mL. At 0.3 mg/mL, all samples gave no test line signal in the VZV LFA. To ensure complete inhibition in all possible samples, 0.5 mg/mL was the gE spike concentration chosen for the cross-reactivity studies. In addition, to show that gE spiking would not interfere with the reactivity of the cross-reactant serum samples, the serum samples were tested with each of the 11 cross-reactant ELISA tests (Table 3).
Table 3. Summary results from cross-reactivity serum samples screened with VZV LFA and Cross-Reactant ELISA.
|VZV LFA Cross-reactivity Study|
|Analyte||Test Results (# Positives /# Tested)|
|Epstein-Barr Virus IgG||0/4|
|Borrelia V1SsE1/pepC10 IgG/IgM||0/6|
|B. burgdorferi IgG||0/6|
|Rheumatoid Factor IgM||0/2|
2.5. Interfering substances
The effect on the detection by VZV LFA of a low level of VZV IgG in a sample containing various potentially interfering substances was evaluated using the following reagents at the highest concentrations that they could potentially be present in a patient whole blood sample: Bilirubin conjugated, bilirubin unconjugated, cholesterol, hemoglobin, triglycerides, human anti-mouse antibodies (HAMA), rheumatoid factor (RF), human serum albumin, alpha interferon, histamine dihydrochloride, glucose, l-ascorbic acid, human IgG and biotin.
One whole blood sample, testing positive for VZV IgG based on the VZV LFA, was aliquoted into 26 samples and half of the samples were spiked with VZV gE antigen to render them “negative”. Samples were prepared and equilibrated at room temperature and tested within 10 min of preparation. All samples were tested with the added substances using two replicates per sample and the results were interpreted for each replicate by two operators to evaluate assay performance.
2.6. Near cut-off study
The objective of this study was to evaluate the performance of the VZV LFA and interpretation of results of samples with concentrations of VZV IgG near the assay cut-off when tested by untrained users at POC facilities and trained users at a reference laboratory setting.
To show performance and results interpretation with VZV IgG samples measuring near the assay cut-off, results from a multi-site reproducibility study were used because the sample panel tested included near cut-off samples and the testing was completed by multiple different users, both trained and untrained. Sample panels were prepared using VZV IgG negative human serum sample pools and positive serum sample pools containing various levels of VZV IgG. Serum pools were mixed to generate a negative, low positive (near cutoff), medium positive, and a high positive signal on the VZV LFA. Each sample was run using four (4) replicates per day at each site.
The goal for producing the low positive or near cutoff sample was to achieve a visual grade between 2 and 4 on the CLIA certified laboratory (DCN) multi color visual grade scale (Fig. 1). This was achieved for all replicates. Three different lots of VZV LFA were used during the reference site study and one lot was used in the multi-site POC study (Table 1). A visual grade <2 is considered a negative sample. The visual color scale was only used for analytical analysis and optimization of VZV LFA. It is not intended to be a tool for end users to quantitate the amount of VZV IgG present in a patient specimen. Many commercial LFA’s are similarly qualitative rather than quantitative tests.
Fig. 1. Multi color visual grade for the detection of VZV LFA test and control lines. A visual grade <2 is considered a negative sample.
Table 1. Samples and Lots used to Qualify Near Cutoff Sample Detectability by End Users.
|Device Lot #||Replicate||Test Line Visual Grade||Avg. Test Line Visual Grade|
- ⁎VZV LFA lots used for the multi-site reproducibility study by untrained users at 2 POC locations.
In addition, the visual color scale was used to corroborate the fingerstick clinical study results performed by untrained operators. A total of 300 fingerstick tests were performed by 9 untrained operators. The results were determined as IgG positive or negative by the operators. All 9 operators were able to correctly determine the fingerstick results without any false positive or negative and without the aid from any visual color scale tool.
2.7. Clinical study design
The clinical study was a prospective, multi-center study designed to determine the performance of the VZV LFA. These centers were one (1) CLIA-certified Reference Laboratory which tested the clinical specimens as serum using the comparator Diamedix ELISA VZV IgG and three (3) POC clinical sites (100 study subjects/site) located in the US. The clinical trial studies were designed and separately performed at 3 outpatient centers for clinical study. Study subjects were enrolled for fingerstick VZV LFA testing and venipuncture blood draw.
All participants provided written informed consent at enrollment. The Study Protocol and Informed Consent Form (ICF) were reviewed and approved by Sterling Institutional Review Board (IRB), Atlanta, Georgia. The study was conducted in accordance with Good Clinical Practice (GCP). Demographic information regarding study participants’ history of childhood chickenpox, and chickenpox and shingles vaccination were also collected.
Study participants included an equal number of males and females, aged 21 years or older. The ethnicity of study subjects is shown in Table 5 and is representative of the US population at large. Inclusion Criteria included: (1) Equal numbers of male and female participants; (2) Age at least 21 years; (3) Patient visit for standard-of-care; and (4) Signed written informed consent. Exclusion Criteria included: (1) Younger than 21 years; and (2) Known to be hemophiliac or anemic.
2.8. VZV LFA and comparator ELISA testing
For fingerstick testing, the VZV LFA cassette was removed from the pouch and placed on a clean, flat surface. A well-beaded drop of blood from the fingerstick puncture site was touched by a capillary tube and the content (15 μL) dispensed into the sample port. Three drops (approx. 100 μL) of chase buffer were added into the sample port. The cassette was incubated at room temperature and the results were read after 20 min.
For comparator ELISA testing, venous whole blood samples were collected in BD Vacutainer blood collection tubes containing clot activator. Serum samples were prepared as described10. Whole blood fingerstick samples were tested by the VZV LFA at the POC sites. The serum samples were tested at the CLIA-certified reference laboratory by both the comparator Diamedix VZV IgG ELISA and the VZV LFA.
2.9. Interpretation of results
Valid Assay. In addition to the presence of the control (C) line, if the test (T) line was also visible, the test result indicated the presence of IgG antibody to VZV gE antigen. An IgG positive (or reactive) result is consistent with previous chickenpox or zoster infection, or successful chickenpox or shingles vaccination. If only the control line (C) is developed, the results indicate that VZV anti-IgG antibody was not detected and is consistent with no previous history of chickenpox or zoster infection or successful vaccination.
Invalid Assay. This test contains a built-in control feature, the control (C) line. The C-line develops after addition of the specimen and sample diluent. If the C-line does not develop, the test is invalid regardless of the color development of the test line (T).
Comparator ELISA and VZV LFA testing characterized specimens as negative or positive for VZV IgG antibody and sensitivity, specificity and positive (PPA) or negative agreement (NPA) performance measurements of the VZV LFA were directly calculated by comparison to the Diamedix ELISA VZV IgG results for the same specimens.
2.10. Measures of test performance
The 19 discordant specimens, and all 279 concordant LFA/Diamedix positive and negative specimens were retested after blocking with purified VZV gE capture antigen to negate any VZV IgG present and to determine the “true IgG positive” status of the clinical samples. The agreement between the VZV LFA and the Diamedix ELISA assay was estimated by assessing positive and negative percent agreement between the two assays. PPA is the proportion of participants who tested positive with the Diamedix assay and who also tested positive in the VZV LFA. NPA is the proportion of participants who tested negative with the Diamedix assay and who also tested negative in the VZV LFA. PPA and NPA with 95% Wilson confidence intervals were calculated for each of the three sites and all sites combined.
2.11. Single-site repeatability study
The objective of this study was to evaluate single-site precision as well as the inter-lot performance of the VZV LFA. This study was designed and executed using input from CLS1 EP05-A3 guideline and was conducted at the CLIA-certified reference laboratory by trained operators.
The within-laboratory repeatability and inter-lot variability of the VZV LFA was analyzed using three different lots of the VZV LFA. The VZV LFA was tested across three days, each of the three lots per day, two different testing events per day (at least 3 h apart), three replicates per sample, by two operators per day. The study was performed using a blinded panel consisting of negative, low-, medium- and high-positive samples. To prepare the panel, two VZV IgG negative human serum sample pools and three positive serum sample pools containing various levels of VZV IgG were mixed to generate negative and low-, medium-, and high-positive visual score signals on the LFA. The pooled sera were randomized, assigned blinded numbers, dispensed into single-use aliquots, and frozen for use in the single-site precision study. To confirm that the serum sample matrix was acceptable for use, refer to the specimen matrix comparison study above that compares whole blood vs. serum matrices.
2.12. Multi-site reproducibility study
The objective of this study was to evaluate the multi-site precision performance of the VZV LFA. This study was designed and executed using CLS1 EP05-A3 guidelines and was conducted at 3 sites: two CLIA-waived POC sites by untrained operators and a CLIA-certified reference laboratory by trained operators.
This study used one VZV LFA verification lot to determine day-to-day, site, and operator precision. Precision testing included testing at three sites, for five days with one testing event per day (time of testing varied day to day) and five replicates per sample.
The study was performed using a blinded panel consisting of negative, and low-, medium-, and high-positive samples. To prepare the panel, two VZV IgG-negative human serum sample pools and three positive serum sample pools containing various levels of VZV IgG were mixed to generate a negative, and low-, medium-, and high-positive signal on the LFA. Pooled sera were randomized, assigned blinded numbers, dispensed into single-use aliquots, and frozen for use in the multi-site precision study.
To confirm that the serum sample matrix is acceptable for use, the specimen matrix comparison study described below was used that compares whole blood, plasma, and serum matrices.
3.1. Specimen matrix comparison
The goal for this study was to compare the performance of VZV LFA when tested using venous whole blood (K2 EDTA), serum and plasma samples. The same donors were tested for all three matrices to compare performance of each on the VZV LFA.
Five (5) individual donors were used and five (5) replicates per matrix condition were tested for each sample. The “blocking” procedure outlined above was not used on any samples for this study. A single factor ANOVA was done to compare the matrices across all samples and showed a p-value ≥0.05. The serum and plasma matrices did however, provide higher visual grade scores overall with one sample exceeding the acceptance criteria of >1.5 visual grade difference. The volume of sample running through the LFA, based on the matrix composition, can explain the differences between the blood and serum/plasma matrices. Since the same volume (15uL) of sample was used for all matrices, this would potentially result in less serum/plasma running through the strip with the blood due to a portion of the total sample volume being composed of hematocrit which gets trapped in the sample pad.
3.2. Laboratory performance
Performance results for the VZV LFA and Diamedix ELISA when testing serial dilutions of the WHO VZV IgG International Standard are shown in Table 2. The results show that visual grades decrease as the WHO VZV IgG standard is further diluted indicating a pattern between the concentration of the WHO standard and the visual grades determined by using the tool in Fig. 1. They indicate that the VZV LFA demonstrates an ability to detect these specific samples at greater dilution factors than the FDA cleared Diamedix ELISA.
Table 2. Comparison of WHO Standard VZV IgG dilution series study results tested with the VZV IgG comparator ELISA test and the VZV LFA.
|Empty Cell||Diamedix ELISA||LFA|
The visual grades were determined using the multi color visual grade tool in Fig. 1.
Cross-reactivity tests were performed using serum samples containing antibodies to the microorganisms and factors listed below in Table 3. Due to the prevalence of IgG to Varicella Zoster Virus (VZV) in the United States, the majority of the population are VZV IgG positive. Similarly, the majority of retrospectively collected commercially available human serum samples are also VZV IgG positive and reactive with the VZV LFA. Therefore, the VZV IgG antibody sites in cross-reactant serum samples were blocked by VZV gE target antigen before the cross-reactivity testing. The results demonstrated that spiking the VZV gE antigen into these serum samples did not inhibit the antibody activity seen in any of the cross-reactant specific ELISAs. On the contrary, the positive signal was completely eliminated in the VZV LFA, indicating that the antigen spike specifically binds to VZV antibody and not to any other potentially circulating cross-reacting antibodies. No samples that were positive changed to negative and no negative samples changed to positive.
3.4. Interfering substances
The effect on the detection by VZV LFA of a low level of VZV IgG in a sample of various potentially interfering substances was evaluated using the following reagents at the highest concentrations that they could potentially be present in a patient’s whole blood sample: Bilirubin conjugated, bilirubin unconjugated, cholesterol, hemoglobin, triglycerides, human anti-mouse antibodies (HAMA), rheumatoid factor (RF), human serum albumin, alpha interferon, histamine dihydrochloride, glucose, l-ascorbic acid, human IgG and biotin. The results indicated that no reagent tested appears to cause interference in VZV LFA’s performance (Table 4).
Table 4. Summary of interfering substance testing of a VZV LFA low-positive and negative result samples.
|Potential Interferent Substance||Added Conc.*||Low-Positive Observed Positive / Expected Positive||Percent Positive Agreement||Negative Observed Negative / Expected Negative||Percent Negative Agreement|
|Bilirubin conjugated||0.4 mg/mL||4/4||100%||4/4||100%|
|Bilirubin unconjugated||0.2 mg/mL||4/4||100%||4/4||100%|
|HAMA (Human anti-Mouse Antibodies)||75 ng/mL||4/4||100%||4/4||100%|
|HSA (Human Serum Albumin)||60 mg/mL||4/4||100%||4/4||100%|
|Histamine Dihydrochloride||4 mg/mL||4/4||100%||4/4||100%|
|Ascorbic acid||0.06 mg/mL||4/4||100%||4/4||100%|
|Human IgG||2 mg/mL||4/4||100%||4/4||100%|
3.5. Demographic and ethnicity analysis of the clinical study samples
The demographic and ethnicity information are shown in Table 5. The study included more females than males. Subject mean age was 45.5 years (SD = 15.7) with a range of 21 to 88. The majority had a history of chickenpox (63%); reported percentages of those that had or had not received a chickenpox vaccination were similar. Only 43 (14.9%) participants reported having had a shingles vaccine. Mean age of those having received the Shingles vaccine was 62.2 (SD = 13.3) with a range of 31 to 88.
Table 5. Demographic, health history and Ethnicity of the 300 at 3 sites (100/site) clinical study participants used to calculate the performance of the VZV LFA.
|Demographic Characteristic||N (%)|
|60 and over||66 (22.15)|
|History of chickenpox||Yes||185 (62.08)|
|History of chickenpox vaccine||Yes||138 (46.31)|
|History of shingles vaccine||Yes||43 (14.43)|
|Ethicistic characteristic||N (%)|
|Black or African American||67 (23.33)|
|American Indian or Alaska Native||3 (1)|
|West Indian||1 (0.33)|
The results from clinical trials showed that 25 VZV LFA Positive study subjects did not have a history of chickenpox, varicella or shingles vaccination; 2 VZV LFA positive had no history of chickenpox and shingles and unknown history of varicella vaccination, suggesting that these subjects may have experienced asymptomatic VZV infection.
In addition, from 13 VZV LFA negatives study subjects, 5 did not have a history of chickenpox, varicella or shingles vaccination; 2 did not have chickenpox or shingles with unknown history of varicella vaccination; 3 had no chickenpox but had varicella vaccination and 3 had chickenpox and varicella vaccination, suggesting that waning VZV antibody immunity in some VZV-infected and/or vaccinated individuals may also result in undetectable level of IgG with VZV LFA as well as ELISA.
3.6. Prospective clinical comparison study
All study participants met the study Clinical Protocol inclusion criteria. A total of 300 participants were enrolled at the three POC sites and fingerstick whole blood samples were prospectively tested at the site with the VZV IgG LFA. Paired, prospectively collected serum samples from the same patients were tested with the VZV IgG ELISA at the Reference Laboratory.
Results from a comparative analysis of fingerstick whole blood samples tested by VZV IgG LFA and paired serum samples tested by the VZV IgG ELISA are shown (Table 6). Agreement between the VZV IgG LFA and the ELISA was estimated by assessing positive and negative percent agreement. Positive percent agreement (PPA) is the proportion of participants who tested positive in the ELISA and in the VZV IgG LFA. Negative percent agreement (NPA) is the proportion of participants who tested negative in the ELISA and in the VZV IgG LFA assay. PPA and NPA with 95% Wilson confidence intervals were calculated for each of the three sites and all sites combined. Results from two invalid LFA results at Site 2 were excluded from final performance calculations Nine equivocal ELISA results were included as “positive” results after spiking /blocking studies indicated that VZV IgG was present in these nine samples.
Table 6. VZV IgG LFA fingerstick data performance against the VZV IgG ELISA.
|Site 1||Comparator Method|
|VZV LFA||Positive (+equivocal)||Negative||Total|
|PPA: % Agreement with Positive Test (95% CI) = 98.89% (93.97, 99.80)|
|NPA: % Agreement with Negative Test (95% CI) = 10.0% (1.79, 40.42)|
|Site 2||Comparator Method|
|VZV LFA||Positive (+equivocal)||Negative||Total|
|PPA: % Agreement with Positive Test (95% CI) = 100% (95.72%, 100%)|
|NPA: % Agreement with Negative Test (95% CI) = 50.0% (25.38%, 74.62%)|
|Site 3||Comparator Method|
|VZV LFA||Positive (+equivocal)||Negative||Total|
|PPA: % Agreement with Positive Test (95% CI) = 98.94% (94.22, 99.81)|
|NPA: % Agreement with Negative Test (95% CI) = 66.67% (30.00, 90.32)|
|All Sites||Comparator Method|
|VZV LFA||Positive (+equivocal)||Negative||Total|
|PPA: % Agreement with Positive Test (95% CI) = 99.26% (97.34, 99.80)|
|NPA: % Agreement with Negative Test (95% CI) = 39.29 (23.57, 57.59)|
The 19 discordant specimens and all 279 concordant LFA / ELISA specimens were retested after adding purified VZV gE capture antigen to block any VZV IgG present to determine the true IgG positive status of the clinical samples.
The number of negative VZV IgG samples obtained during the clinical study was low (13/300; 4.3%) due to a large percentage of the US population being vaccinated for or naturally infected with VZV at some time during their lives. Attempts were made to acquire a larger number of retrospectively obtained VZV seronegative serum samples from US suppliers (LGC Clinical Diagnostics, Discovery Life Sciences, NOVA Biologics, USBiolab), and from suppliers of specimens from outside the US (Precisions for Medicine, Boca Biolistics, ZeptoMetrix) as well as WHO, Georgia and California State Health Departments. They all indicated that either they did not have any VZV IgG negative serum samples in their inventory or they do not collect VZV specimens. This low prevalence of negative samples in the VZV IgG LFA clinical trial, further corroborated that much of the US population has VZV immunity. Sourcing specimens from outside of the US was difficult and excessively burdensome these past two years due to COVID’s disruption of worldwide healthcare systems.
3.7. Discordant analysis blocking study results
To block VZV gE antibody sites, study subjects’ serum samples were pre-adsorbed with gE antigen and tested with the VZV IgG LFA. The true PPA and NPA of VZV IgG LFA was determined by testing all 298 clinical study samples including the discordant (19) and all concordant negative and positive (279) study serum samples before and after addition of VZV IgG LFA gE antigen. These studies demonstrated that gE addition was capable of blocking IgG antibody sites in all 287 positive and 11 negative clinical serum samples tested, resulting in 298 VZV IgG LFA IgG negative results. The results also showed that all 19 discordant LFA/comparator specimens contained VZV IgG antibody, resulting in a recalculated performance for VZV IgG LFA in this blocking study of 100% positive and negative agreement for the detection of VZV IgG antibody in all 298 serum clinical study samples.
However, the final percent agreement of the LFA for the fingerstick whole blood samples demonstrated to contain VZV IgG was 99.30% (285/287) (95% CI 97.50–99.81) whereas, the NPA improved from 39.29% (23.57–57.59) to 100% (95% CI: 74.12–100). The final percent agreement of the predicate ELISA for serum samples demonstrated to contain VZV IgG was 94.08% (270/287) (95% CI 90.72–96.27). The accuracy of the VZV IgG LFA test was 99.33% (95% CI 97.59–99.82). The accuracy of the comparator was 93.62% (95% CI 90.26, 95.88).
3.8. Near cut-off study
The results from this study (Table 7) showed that expected result interpretations for near cut-off sample replicates were consistent for both untrained and trained users at all three sites and across all five (5) days of testing.
Table 7. Results from Three Different Sites Testing Near to Cut-off Samples.
|Empty Cell||Empty Cell||VZV LFA Interpretation|
|Site||Sample #||Day 1||Day 2||Day 3||Day 4||Day 5|
|POC Clinical Site 1||1||Positive||Positive||Positive||Positive||Positive|
|POC Clinical Site 2||1||Positive||Positive||Positive||Positive||Positive|
In addition, results from visual characterization of clinical fingerstick samples negative by VZV LFA testing and clinical serum samples negative by VZV LFA and ELISA testing (Table 8) demonstrated that untrained users detected as positive not only close to the cut-off samples, but also more moderate positives much better than the comparator ELISA.
Table 8. Visual read of results obtained for testing by VZV LFA (fingerstick), LFA (serum) and ELISA (serum).
|Sample ID||LFA POC Clinical Sites Visual Read Only Characterization Fingerstick||LFA Reference Laboratory Site Visual Grade Characterization (Average of 3 serum reps)||ELISA Results Characterization|
Note: As in this near-cutoff test, when a patient specimen has a very low level of VZV IgG it can produce a positive result when the LFA is run on serum but the same patient’s fingerstick whole blood sample will be negative. In the Sample Matrix study, serum and plasma samples provided slightly higher visual grade scores overall, as expected. The volume of sample running through an LFA cassette, based on the matrix composition, can explain the differences between the blood and serum or plasma matrices. Since the same 15 μL volume of sample was used for all sample types, this would typically result in a lower blood sample (plasma) volume running through the strip due to a portion of the total sample volume being composed of the cellular components (hematocrit, 37–52% normal range) which gets trapped in the sample pad.
3.9. Single-site repeatability study
There was 99.6% agreement for observed vs. expected result interpretation for the positive and negative samples across all operators, days, device lots and events (Table 9 and Table 10). The one run that was discordant showed an acceptable test line signal but was deemed invalid due to a control line signal below the visual grade cut-off. The attribute agreement analysis across the three lots showed kappa values of >0.95.
Table 9. Results from the singe-site repeatability study by lot.
|Sample VZV IgG Level||Lot 1% (n/N)||Lot 2% (n/N)||Lot 3% (n/N)||3 Lots Combined % (n/N)||3 Lots Combined 95% CI|
|High||100% (18/18)||100% (18/18)||100% (18/18)||100% (54/54)||93.4%, 100%|
|Med||100% (18/18)||100% (18/18)||100% (18/18)||100% (54/54)||93.4%, 100%|
|Low||100% (18/18)||94.4% 17/18)1||100% (18/18)||98.1% (53/54) (1)||90.2%, 99.7%|
|Neg 1||100% (18/18)||100% (18/18)||100% (18/18)||100% (54/54)||93.4%, 100%|
|Neg 2||100% (18/18)||100% (18/18)||100% (18/18)||100% (54/54)||93.4%, 100%|
|Overall||100% (90/90)||98.9% (89/90)||100% (90/90)||99.6 (269/270)||97.9%, 99.9%|
One replicate had an invalid result.
Table 10. Single-Site Repeatability Results by Operator.
|VZV IgG Level||Op 1% (n/N)||Op 2% (n/N)||2 Ops Combined % (n/N)||2 Ops Combined 95% CI|
|High||100% (27/27)||100% (27/27)||100% (54/54)||93.4%, 100%|
|Med||100% (27/27)||100% (27/27)||100% (54/54)||93.4%, 100%|
|Low||100% (27/27)||96.3% (26/27)||98.1% (53/54)||90.1%, 100%|
|Neg 1||100% (27/27)||100% (27/27)||100% (54/54)||93.4%, 100%|
|Neg 2||100% (27/27)||100% (27/27)||100% (54/54)||93.4%, 100%|
Attribute agreement analysis across three (3) lots showed kappa values of at least 0.97 exceeding the pre-set acceptance criterion of 0.80.
Attribute agreement analysis across two operators showed kappa values of at least 0.98, exceeding the pre-set acceptance criterion of 80%.
3.10. Multi-site reproducibility study
There was 100% negative agreement between all sites and days and a 99.7% positive agreement between all sites, operators and days. Attribute agreement analysis showed a kappa value of 0.99 between sites (Table 11).
Table 11. Multi-site reproducibility study results.1
|VZV IgG Level||Site 1% (n/N)||Site 2% (n/N)||Site 3% (n/N)||3 Sites Combined % (n/N)||3 Sites Combined 95% CI|
|High||100% (25/25)||96.0% (24/25) 1||100% (25/25)||98.7% (74/75)||92.8%, 99.8%|
|Med||100% (25/25)||100% (25/25)||100% (25/25)||100% (75/75)||95.1%, 100%|
|Low||100% (25/25)||100% (25/25)||100% (25/25)||100% (75/75)||95.1%, 100%|
|Neg 1||100% (25/25)||100% (25/25)||100% (25/25)||100% (75/75)||95.1%, 100%|
|Neg 2||100% (25/25)||100% (25/25)||100% (25/25)||100% (75/75)||95.1%, 100%|
|Overall||100% (125/125)||99.2% (124/125)||100% (125/125)||99.7% (374/375)||98.5%, 100%|
The one interpretation that did not match came from a high positive sample run that was interpreted as a negative result. This run was valid as a control line was said to be present and visible by the operator, so it is unclear what caused the erroneous result.
Attribute agreement analysis across the three (3) sites showed kappa values of at least 0.98 exceeding the pre-set acceptance criterion of 0.80.
A single serologic VZV IgG test can be used to determine if a person has antibodies to VZV from past varicella disease or who may be candidates for chickenpox or shingles vaccination. ELISA tests can readily detect seroconversion to natural infection with VZV. However, commercially available VZV IgG assays are not sensitive enough to detect all seroconversions after vaccination (CDC Chickenpox (Varicella), 2021).
While many commercial VZV IgG ELISAs perform well enough to reliably detect seroconversion for infection by wild type virus, the performance specifications (specificity/sensitivity or PPA/NPA) of these methods vary widely. Some commercially available VZV IgG assays have been found to be unreliable, even for the detection of VZV IgG after vaccination (CDC Chickenpox (Varicella), 2021). This is due to the large number of viral structural, non-structural and envelope glycoproteins used as capture antigens in commercial ELISA tests. VZV glycoproteins as well as other VZV proteins containing antigenic epitopes will elicit antibodies after VZV infection or vaccination. This will result in a large number of circulating VZV antibodies in a patient’s blood that are not representative of true immunity to VZV infection.
Unlike an ELISA test, the VZV LFA test line consists of only a highly purified truncated VZV recombinant gE protein. VZV gE is the most abundant and immunogenic VZV envelope glycoprotein capable of inducing both humoral and cell-mediated immunity and is used in an adjuvanted subunit vaccine (SHINGRIX®) to prevent zoster or shingles in individuals aged 50 years and older.
Current commercially available ELISA tests require collection, transport and processing of blood samples to separate serum or plasma for further testing. The VZV LFA is used for the detection of IgG antibodies against VZV by testing patient whole blood samples using a lateral flow immunoassay (LFA) test. The VZV LFA does not require collection and processing of blood samples and provides a simple and rapid POC test for the detection of VZV antibody in whole blood from a fingerstick sample.
The VZV LFA is similar to the Diamedix ELISA in several respects. Like the Diamedix ELISA, the VZV LFA use is to detect IgG antibodies to the varicella zoster virus and can be used as an aid in the determination of a previous chickenpox infection. In addition, the VZV LFA has been shown to aid in the determination of a previous zoster infection, or successful chickenpox or shingles vaccination. Both tests can be used in CLIA-certified moderate and high complexity clinical laboratories by trained medical technologist professionals, but the VZV LFA has also been demonstrated to be simple and sufficiently robust to be used in CLIA-waived POC facilities by untrained healthcare professionals. The analytical performance of the VZV LFA is comparable to or better than the comparator ELISA and the clinical study shows substantially equivalent or better results for the VZV LFA. Both tests are immunology capture antigen/VZV IgG-based but differ in their methodology. The VZV LFA is a rapid, lateral flow chromatographic immunoassay using a cartridge with a nitrocellulose membrane as the solid phase whereas, the Diamedix test is an enzyme-linked immunosorbent assay (ELISA) using a Polystyrene 96 well plate as the solid phase. The VZV LFA employs a capture antigen that is a purified truncated recombinant VZV glycoprotein E (gE) and the comparator test uses partially purified VZV viral lysate proteins. The specimen type for the VZV LFA is whole blood fingerstick and for the ELISA test it is serum.
One of the significant differences between the VZV LFA and the predicate device is the method of detection. The ELISA test uses a spectrophotometer to measure optical density of a color reaction and calculates a semi-quantitative result within 90 min, whereas the VZV LFA is visually read as a qualitative positive or negative result within 20 min by the operator. While all previously FDA cleared ELISA VZV IgG tests use instrumentation to read the results of the test, this test would be the first CLIA-waived rapid VZV IgG diagnostic test lateral flow immunoassay, which in addition is also visually read. Untrained operators in physicians’ offices and other POC locations are accustomed to reading these types of LFAs. The VZV LFA was used at three intended user POC sites and untrained users successfully read and interpreted the results of the test, even for specimens with low levels of VZV IgG. The VZV LFA was demonstrated to have improved sensitivity/PPA and specificity/NPA over the comparator ELISA.
Clinical performance characteristics of VZV LFA IgG fingerstick whole did not show any false positive results as determined when VZV IgG binding sites in clinical serum samples obtained from the 300 subjects were blocked by spiking with gE antigen. In addition, as determined by analytical results and clinical testing at POC facilities by untrained users, the risk of false negative results with the VZV LFA is also low compared with the comparator ELISA test. In addition, an invalid result can be repeated at the clinic and will not result in a delay in diagnosis and extra expense to a user for the recommendation to repeat a blood draw and serum sample test.
With low LFA false positive and false negative results, patients may neither receive an appropriate vaccination, be potentially given unnecessary treatment or later develop chickenpox or shingles as is presumed immune by a false positive result.
A POC VZV LFA will help avoid risk of a delay in both the diagnosis, vaccination and treatment of chickenpox infection or shingles due to VZV reactivation. It will also help management and prevention of varicella in the seronegative expected mothers.
The performance of the VZV LFA in both the laboratory and clinical field evaluations showed the device to be a sensitive and specific VZV IgG detection system that is easy to use, allows a quick and consistent reading and interpretation of results, and is capable of being used in both laboratory and non-laboratory environments.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
We thank Cindy Clark for assistance with reviewing, editing, and for comments that greatly improved the manuscript.
Data will be made available on request.