Use of vaccines against the SARS-CoV-2 virus has been a primary strategy to prevent COVID-19 infection1. Neutralizing antibody titers have been found to be strongly correlated with observed vaccine effectiveness against symptomatic and severe COVID-192. Indeed, both the initial Moderna and Pfzer/BioNtech BNT162b2 vaccines elicited strong IgG immune responses, with neutralizing antibodies considered to be a significant part of the vaccine-related protection against COVID-193. Unfortunately, the effectiveness of vaccine protection and neutralizing antibody levels appears to decline in the first six months or so after vaccination4. Quantifying vaccine-related immune responses to SARS-CoV-2 may therefore have potential clinical utility in determining personal immunity against COVID-19 as part of individualized vaccination or treatment decisions. However, antibody testing is not currently recommended according to current CDC guidelines5.

Despite this recommendation, many methods have been developed to measure neutralizing antibodies to SARS-CoV-2, including the plaque reduction neutralization test, considered a reference standard6, pseudovirus neutralization assays, ELISAs (enzyme-linked immunosorbent assays), and a number of lateral flow immunoassays7,8,9,10,11,12,13,14. Most of the lateral flow assays that measure neutralizing antibodies are based on inhibition of the interaction of the Spike protein receptor binding domain and the cellular angiotensin converting enzyme 2 (ACE2) receptor. Neutralizing antibodies can prevent viral infection by interfering with viral binding and cellular internalization of the virus. Detecting the presence of neutralizing antibodies may indicate prior infection with SARS-CoV-2 or previous immunization with a COVID-19 vaccine. Despite the number of assays reported, of the 83 serology assays tests that have received EUA from the United States Food and Drug Administration (FDA)15, only three were neutralizing antibody assays (2 ELISAs and one chemiluminescent immunoassay). All of the FDA EUA authorized lateral flow immunoassays measure total antibody binding to SARS-CoV-2 antigens. Both of the FDA EUA authorized neutralizing antibody ELISAs were approved as high complexity tests. Such tests require specialized equipment and reagents, highly trained personnel, and would likely require that samples be transported to a laboratory for testing, creating potentially long delays between sample acquisition and results.

We evaluated the performance of a lateral flow immunochromatography assay measurement system for detecting SARS-CoV-2 neutralizing antibodies against the FDA EUA authorized SARS-CoV-2 Neutralization Antibody Detection cPass™ Kit (Genscript cPass neutralization antibody assay, Piscataway, NJ, USA)16,17,18. A previous study found inverse levels of sensitivity and specificity for different thresholds of detection using a lateral flow assay for neutralizing antibodies with detector/reader instrument8. Other data suggest that manually read lateral flow assays had poor sensitivity after the initial COVID-19 vaccination13 and a fluorescent neutralizing antibody lateral flow assays with reader had a correlation r2 of 0.2 with ELISA results14. We used a portable, light-weight spectrometer to provide a semi-quantitative measurement of neutralizing antibody levels, similar to other methods to quantify lateral flow devices11, but on a large population of individuals on whom both vaccination and RNA testing status were known19, though larger studies of vaccinated populations lacking infection status have also been reported8,9. Our results indicate that the lateral flow spectrometer system had positive predictive and negative predictive values comparable to ELISA for the measurement of neutralizing antibodies in SARS-CoV-2 vaccinated and previously infected individuals. Such a system may be particularly useful for studies of vaccination response in which access to ELISA methodology is not available or practical.

Materials and methods

Study design and participants

We conducted a single-center retrospective cross-sectional study at the Temple University Health System in Philadelphia, PA. Samples were obtained from clinically ordered EDTA anti-coagulated whole blood samples analyzed by the TUHS Clinical Laboratory and stored for up to three days prior to discard collected from May 2021 through 2022. Only de-identified samples and data were used. The study was conducted in accordance with relevant guidelines and regulations as required by the Institutional Review Board of Temple University who approved the study protocol. Research ethics approval was obtained from the Temple University Institutional Review Board (IRB No. 29810 and 25673). The need for informed consent was waived by the Institutional Review Board of Temple University.

Participants were included if (1) they were at least 18 years old, (2) had a blood sample sent to the TUHS clinical laboratory, and (3) had data in the electronic health record. Exclusion criteria were (1) adequate volume of plasma obtained for testing, (2) visually identified hemolysis, and (3) visually identified lipemia.

Samples

Plasma samples were obtained from K2EDTA anti-coagulated whole blood. Blood tubes were first inverted 5 times then placed on ice followed by centrifugation at 1200×g for 10 min at 4 °C. Aliquots of plasma from the upper plasma layer were collected and stored at − 80 °C until analysis. Batches of consecutive samples were randomly selected. For the comparator study, clinical information, ELISA, and DiaVac-Spectroclip SARS-CoV-2 Neutralizing Antibody Test System results were not known to the personnel conducting the assays.

Neutralizing antibody lateral flow-spectrometer assay

A commercially available SARS-CoV-2 Neutralizing Antibody Test Kit (Healgen Scientific—LFIA test kit, Healgen Scientific, LLC, Houston, USA) based on lateral flow immunochromatography was used to detect the presence of neutralizing antibodies along with colorimetric spectrophotometry20,21,22. A recombinant SARS-CoV-2 spike protein S1 subunit receptor binding domain (RBD) peptide and a tag protein were conjugated to gold nanoparticles (GNP) and embedded in the conjugate pad of the lateral flow chromatography strip of nitrocellulose membrane. The SARS-CoV-2 virus Spike (S) protein S1 RBD binds to the cell surface receptor Angiotensin-Converting Enzyme 2 (ACE2) present on the surface of epithelial cells, causing infection of human cells. Recombinant human ACE2 protein (Gln 18-Ser740 produced in a Chinese Hamster Ovary cell line confirmed by mass spectrometry) was immobilized on the nitrocellulose membrane to form Test (T) line and an anti-tag antibody was immobilized on the nitrocellulose membrane to form a Control (C) line. Goat anti-mouse IgG served as the anti-tag antibody immobilized on the membrane at the C line. The tag protein was mouse IgG that was also conjugated with gold nanoparticles and adsorbed to the nitrocellulose strip in the area between the sample addition spot (S) and the T line. When the fluid sample migrated by capillary action through this area, the mouse IgG conjugated with gold nanoparticles was transported past the T line to the C line region where it is bound by the immobilized goat anti-mouse IgG to form a visible line. This C line serves as a control for a successful lateral flow movement of sample.

When an adequate volume of test specimen is dispensed into the sample well of the test cassette, the specimen migrates by capillary action past the conjugate pad along the cassette across the T and C lines. The S1-RBD conjugated to GNP is captured at the T line by the immobilized ACE2 protein during capillary transport along the nitrocellulose membrane. When neutralizing antibodies exist in the test sample, they will bind to the S-RBD in the GNP prior to and during transport to the T line, blocking or neutralizing the interaction between S1-RBD and ACE2. Blocking this interaction leads to a reduction or complete loss of the signal at the T line, depending on the level of neutralizing antibodies in the test sample (Fig. 1). In contrast, if no neutralizing antibodies are present in the test sample, there will be no interference with the interaction between the S1-RBD in the GNP and the ACE2 immobilized at the T line resulting in full colorimetric development of the line. The immobilized anti-tag antibody at the C line will capture tag protein of the GNP to show a colored band, serving as a built-in positive control that capillary action brought the sample fluid and the necessary reagents along the chromatography strip.

Fig. 1
figure 1

Principle of neutralizing antibody detection at the test (T) line.

Full size image

Prior to testing, the test cassette, whole blood, serum, or plasma specimens, and buffer were allowed to equilibrate to room temperature. After placing the test cassette on a clean and level surface, approximately 25 μL of serum or plasma was added to the specimen well of the test cassette for absorption and capillary action flow through the chromatography strip. For whole blood, approximately 50 μL of sample is added followed by approximately 40 μL of buffer. After 10 min to allow full transit of the sample through the entire length of the chromatography strip, the cassette was inserted onto the slot/carrier of the DiaVac-Spectroclip spectrophotometer device (Supplementary Methods M1) (Fig. 2). The door was closed and the test button in the computer operating software was pressed to begin scanning. The spectrophotometer device first conducted a position calibration procedure. Because the spectrophotometer device used a stepper motor without an encoder, optical electric switches on both sides of the device were used. When the position calibration procedure began, the slot/carrier with cassette moved toward one side, triggering the switch on that side. The slot/carrier with cassette then moved to the other side, triggering the switch on that side. After this procedure, the software has ascertained the position of the carrier for the test reading.

Fig. 2
figure 2

Device operation and measurement. Plasma, serum, or whole blood samples are added at the S location on the lateral flow cassette. After 10 min, the lateral flow cassette is inserted into a miniaturized spectrometer (DiaVac SpectroClip®) for spectroscopic measurements. The spectroscopic measurements are sent through a cable connection to a personal computer for analysis by a software algorithm that generates a semi-quantitative result that is displayed on the PC.

Full size image

Following ascertainment of cassette position, the spectrophotometer device then scanned the C line and T line of the cassette test strip. The spectrum-based reader provides high-resolution (3–5 nm) results across a wide spectral range (300–1100 nm). The primary reflectance wavelengths detected using the spectrum reader were 500 nm and 600 nm, with a main reference wavelength of 680 nm. The positions on the test strip of the C line and T line were mapped using data collected from measuring the positions of the C line and T line of hundreds of test strips. The device will thus first scan the C line starting from a predefined point that will be outside the area of the C line. The device will scan the reflectance at 9 points on the cassette test strip crossing over the position of the C line. Each point is a circle with a 0.6 mm diameter. When the device scans one point, the slot/carrier is moved by the motor along the length of the cassette test strip by 0.15 mm. The platform will move a total of 1.35 mm across the 9 points to encompass the widths of the C and T lines.

For each point scanned, a reflectance spectrum is acquired. The single point measurement indicating the most intense C line color development is used in the calculation of the neutralizing antibody level (below). The system then moves the slot/carrier to the T-line position. The spectrum of the cassette strip between the C and T lines, but nearer to the C line, is also obtained to generate a control or background spectrum. The platform then scans 9 points to capture the single point with maximum reflection spectrum for the T-line. After testing the T-line, the device will move to the white area near the T-line again to obtain the reflectance spectrum near the T-line to calculate the control or background spectrum for the T-line. Once the device has acquired the reflectance of the C line and T line it will convert the reflectance to absorbance, which will be used to calculate the concentration of the neutralizing antibody.

The reflectance results were log- transformed to absorbance via the formula Absorbance =  − log(reflectance).

$${N}_{ab}=1-\frac{{({A}_{\text{max}(500\sim 600 nm)}-{A}_{680 nm})}_{Test line}}{{({A}_{\text{max}(500\sim 600 nm)}-{A}_{680 nm})}_{Control line}}$$

The DiaVac-Spectroclip Spectrometer interprets the results using a semi-quantitative scoring algorithm, and reports a POSITIVE, POSITIVE MAX, NEGATIVE or INVALID result after approximately 150 s. The test results were displayed on software running on Windows 10.

Neutralizing antibody ELISA

The FDA EUA authorized GenScript cPASS SARS-CoV-2 Neutralization Antibody Detection kit (cPASS; GenScript, USA Inc., Piscataway, NJ, USA), a competitive enzyme-linked immunosorbent assay that detects NAb against SARS-CoV-2 that has been used as a standard in other studies23, was performed according to the manufacturer’s protocol. Briefly, plasma samples and kit controls were diluted with kit buffer and run in duplicate. Neutralization reactions were prepared by mixing the positive and negative controls and plasma samples with horseradish peroxidase-labeled RBD solution. Neutralization reactions were then added to a 96-well plate that was precoated with recombinant angiotensin-converting enzyme 2 (ACE2) receptor protein and incubated at room temperature for 25 min. The plate was then washed four times with wash solution followed by addition of 3, 3′, 5, 5′-Tetramethylbenzidine solution to each well with incubation in the dark for 15 min at 25 °C. Stop solution was then added followed by measurement of the optical density (OD) at 450 nm (Tecan Sunrise™). The percentage of inhibition (% inhibition) from neutralizing antibodies was calculated as: % inhibition = (1 − (OD value of sample/OD value of negative control) ) × 100. Results with % inhibition ≥ 40% were considered as positive and those with % inhibition < 40% were considered negative (Table S1). Samples with 100% inhibition were classified as positive max.

Performance characteristics and statistical analyses

The limit of detection (LoD) was determined using contrived plasma specimens containing stock SARS-CoV-2 Neutralizing Antibody Calibrator (GenScript #A02087). The SARS-CoV-2 Neutralizing Antibody (NAb) Calibrator was supplied in a stock solution at a concentration of 1 × 106 Units (U) per mL (U/mL). A diluted stock solution of 6,000U/mL was prepared by mixing 6μL of the Stock with 994μL of the kit supplied Sample Dilution Buffer. This was then diluted into ELISA SARS-CoV-2 neutralizing antibody negative plasma using 20 replicates run at three different NAb calibrator concentrations.

Repeatability was calculated using percent inhibition values. The “within lab” metric was the standard deviation (SD) divided by square root of the number of samples N, the “between days” metric was calculated using the SDs of the means of day 1, day 2, and day 3, and the “between runs” metric was calculated using the average of the SDs of day 1, day 2, and day 3. The coefficient of variation (CV) was calculated using SD/Mean.

Interference assays were performed by preparing stock solutions with 5X concentration of the desired interfering substance using molecular grade HPLC water. Five uL of the interfering substance stock solution was then added to 20 uL plasma sample and mixed immediately prior to loading in to the sample well of the test cassette.

Cross-reactivity analysis was performed using human plasma samples from patients exposed to Influenza A (IgG and IgM), Influenza B (IgG and IgM), hCoV229E (IgG and IgM), hCoV OC43 (IgG and IgM), SARS-CoV-1, and MERS-CoV IgG. In addition, pooled plasma that was negative by ELISA for neutralizing antibodies was spiked with antibodies raised against RSV, Varicella Zoster Virus, Influenza B Virus, Avian Influenza H5N1, HCoV-OC43, HCoV-NL63, HCoV-229E, Mumps Virus, Rhinovirus VP3, Chlamydia Pneumoniae, Haemophilus Influenza B, H7N9, Enterovirus 71, Influenza A H1N1, Mycoplasma Pneumoniae, Measles Virus, Influenza A H3N2, Adenovirus, Rotavirus, Parainfluenza Virus Type 2/3, and HCoV-HKU1.

Sensitivity was calculated as: [a/(a + c)] × 100. Specificity was calculated as: [d/(b + d)] × 100. Positive predictive value (PPV) was calculated as [a/(a + b)] × 100. Negative predictive value (NPV) was calculated as: [d/(c + d)] × 100, where:

$$\begin{array}{*{20}l} {} & {{\text{DIAVAC POS}}} & {{\text{DIAVAC NEG }}\left( { < {\text{4}}0\% } \right)} \\ {{\text{cPASSPOS}}} & {\text{a}} & {\text{b}} \\ {{\text{cPASSNEG }}\left( { < {\text{4}}0\% } \right)} & {\text{c}} & {\text{d}} \\ \end{array}$$

Linearity plot confidence intervals were calculated using GraphPad Prism 10.0. The 95% confidence intervals (CI) for comparator performance studies were calculated using the Score Method (Wilson). Compliance with STARD guidelines is shown in Table S2.

Results

Limit of detection

We first determined the limit of detection using contrived plasma specimens in which ELISA SARS-CoV-2 neutralizing antibody negative plasma containing SARS-CoV-2 Neutralizing Antibody Calibrator was assayed using 20 replicates run at three different neutralizing antibody calibrator concentrations (data shown in Table S3). The percentage of 20 replicates that was positive (≥ 40%) for 30 U/ml was 70%, the percentage of 20 replicates that was positive (≥ 40%) for 35 U/ml was 85%, and the percentage of 20 replicates that was positive (≥ 40%) for 40 U/ml was 95%. The LOD was therefore designated as 40 U/mL, determined as the concentration with proportions of false positives less than 5% and false negatives less than 5% within the linear range of the test.

Linearity

Linearity testing was performed using the same stock SARS-CoV-2 Neutralizing Antibody Calibrator from 37.5 to 600 U/ml. Linearity (r2 = 0.9865) was demonstrated throughout this measuring interval between the DiaVac-Spectroclip SARS-CoV-2 Neutralizing Antibody Test System and the calibrator concentration (Fig. 3 with data shown in Table S4).

Fig. 3
figure 3

Linearity plot using dilutions of SARS-CoV-2 Neutralizing Antibody Calibrator from 37.5 to 600 U/ml.

Full size image

Repeatability

We then determined how repeatable results were using five plasma samples: two positive max samples with 100% inhibition (neutralization), two positive samples (one intermediate and one weak positive), and 1 negative sample. Each was tested using 8 replicates on each of three consecutive days to calculate repeatability, “between runs”, “between days”, and “within lab” precision (data shown in Table S5). The coefficient of variation of percent inhibition for repeatability, “between runs”, “between days”, and “within lab” results was < 5% for two strong positive samples, < 7.02% for intermediate positive results, < 16.23% for weak positive results (Table 1). All negative samples tested as negative.

Table 1 Repeatability, between runs, between days, and within lab results.
Full size table

Interference

To evaluate the effect of common potentially interfering substances found in blood on the performance of the DiaVac-Spectroclip SARS-CoV-2 Neutralizing Antibody Test System, each potential interferent was added to one sample negative for SARS-CoV-2 neutralizing antibodies and one sample positive max for SARS-CoV-2 neutralizing antibodies. None of the substances caused an increase in neutralizing activity in the negative sample, and none caused a decrease in percent neutralization in the positive max sample that would result in a negative result (Table 2), similar to results for the same interfering substances at the same concentrations for an FDA EUA authorized SARS-CoV-2 antibody lateral flow device24.

Table 2 Impact of common potentially interfering substances on negative and positive max samples.
Full size table

Cross-reactivity (analytical specificity)

Serum samples were obtained from patients with natural immunity to the viruses listed in Table S6 to test for potentially cross-reacting sub-groups that could cause false positive indications of neutralizing antibodies. Two individuals for each antibody were analyzed: hCoV OC43 (2 with IgG and 2 with IgM), hCoV229E (2 with IgG and 2 with IgM), Influenza A (2 with IgG and 2 with IgM), Influenza B (2 with IgG and 2 with IgM), MERS-CoV (2 with IgG), and SARS-CoV-1. Of these, 3/4 samples from MERS-CoV IgG were found to be positive (≥ 40% threshold) using the DiaVac SARS-CoV-2 Neutralization Antibody Detection Strip and Reader test (data for manufacturer, catalog number, lot number, percent inhibition, and interpretation shown in Table S6).

ELISA SARS-CoV-2 neutralizing antibody negative plasma samples spiked with commercially obtained antibodies raised against RSV, Varicella Zoster Virus, Influenza B Virus, Avian Influenza H5N1, HCoV-OC43, HCoV-NL63, HCoV-229E, Mumps Virus, Rhinovirus VP3, Chlamydia Pneumoniae, Haemophilus Influenza B, H7N9, Enterovirus 71, Influenza A H1N1, Mycoplasma Pneumoniae, Measles Virus, Influenza A H3N2, Adenovirus, Rotavirus, Parainfluenza Virus Type 2/3, and HCoV-HKU1 were also tested for potential neutralizing activity (information on clonality, manufacturer, catalog number, lot number, concentration, number of runs, and results are shown in Table S7). None were found to be reactive against the DiaVac SARS-CoV-2 Neutralization Antibody Detection Strip and Reader test.

Matrix equivalency

Matrix equivalency studies were conducted to evaluate the performance of whole blood versus EDTA anticoagulated plasma. In these studies, EDTA anticoagulated whole blood was tested in parallel with plasma obtained from the same whole blood samples. All results for plasma and whole blood samples from 32 ELISA SARS-CoV-2 neutralizing antibody positive and five ELISA SARS-CoV-2 neutralizing antibody negative individuals were concordant with each other, and the results were also concordant with ELISA analysis of the plasma from these individuals (data shown in Table S8).

Comparator analysis

The comparator method used to evaluate clinical performance was the FDA EUA authorized Genscript cPass SARS-CoV-2 Neutralization Antibody Detection Kit Blocking Enzyme-Linked Immunosorbent Assay (ELISA) intended for the qualitative and semi-quantitative direct detection of total neutralizing antibodies to SARS-CoV-2 in human serum and EDTA anticoagulated plasma. Information of SARS-CoV-2 infection history and vaccination status was obtained from the electronic medical record. As presented in Table 3 (data and calculations shown in Table S9), the results for 274 plasma samples showed a positive predictive value (PPV) of 99.0% (CI 96.4–99.7%) and a negative predictive value (NPV) of 91.9% (CI 83.4–96.2%) for the DiaVac SARS-CoV-2 Neutralization Antibody Detection Strip and Reader compared to ELISA. A plot of the results (%) for each method is shown in Figure S1.

Table 3 Positive predictive value (PPV) and negative predictive value (NPV) results for all samples and for vaccine subgroups.
Full size table

We then evaluated the performance of the DiaVac SARS-CoV-2 Neutralization Antibody Detection Strip and Reader in subgroups based on prior history of PCR positive SARS-CoV-2 diagnosis and vaccine type and number (data for subgroups shown in Table S9 and results for subgroups in Table 3). The PPV for all subgroups was > 95% except for those individuals who had only one Pfizer vaccination (PPV of 80%). The NPV for those who were PCR positive with no vaccinations was 100% while only 88.1% for those without a previous positive test or vaccination. The NPV for those with Pfizer vaccinations was 80% in contrast to 100% for those with Moderna vaccinations. We also considered age and time interval since vaccination, S11, and S12) for those who received one Janssen (data in Tables S10) and up to three Moderna (data in Table S11) or Pfizer vaccinations (data in Table S12). No clear trends with age, vaccine number, or post-vaccine time interval were apparent except for the expected low antibody responses within 15 days of the first Pfizer vaccine.

Discussion

Our main objectives were to characterize the performance characteristics of a lateral flow-spectrophotometer device requiring a small volume of plasma and determine whether the diagnostic performance was comparable to an FDA EUA authorized ELISA method for the measurement of SARS-CoV-2 neutralizing antibodies. The availability of data on SARS-CoV-2 infection and vaccination also allowed us to evaluate test system in the setting of different vaccination scenarios. We found that the system limit of detection, linearity, repeatability, interfering substance, and cross-reactivity studies to be robust. In cross-reactivity studies, only MERS-CoV IgG were found to be reactive with the DiaVac SARS-CoV-2 Neutralization System. Our comparator study of 274 plasma samples showed a positive predictive value of 99.0% and a negative predictive value of 91.9% compared to one of only two ELISAs for neutralizing antibodies granted FDA EUA status.

Assays that detect the presence of antibodies with specificity for specific viral proteins include ELISA, chemiluminescence immunoassays, lateral flow assays, and neutralization assays25. Plaque reduction neutralization assays have been regarded as the reference method for detecting of neutralizing antibodies, but require the infection of cells in culture by live virus, while pseudovirus neutralization assays also require live cells and a synthetic virus26. ELISA and chemiluminescent assays are more sensitive than lateral flow assays, but require specialized equipment and technical skill and/or automation to perform.

In contrast, lateral flow-based assays have been widely used as point-of-care tests because they require less skill and resources to perform. Soon after the onset of the COVID-19 pandemic, numerous ELISA and lateral flow assays for rapid antibody detection were granted emergency use authorization27. Most of these assays were not specific for neutralizing antibodies, though may be an indirect measure of neutralization28. In contrast, our lateral flow assay detected neutralizing antibodies. It detected the loss of signal from S1-RBD-ACE2 binding, more difficult to quantify by human visual observation relative to the appearance of a colored line. The use of a small, portable, easy-to-use spectrophotometer allowed for semi-quantitative measurement of the lateral flow test strip that had a 99% PPV relative to ELISA. The lower NPV (91.9%) is likely due to inherent variability in the strength of color development of the control reaction and test reaction lines resulting in a low number of false negative results.

The clinical utility of antibody-based SARS-CoV-2 neutralization assays is to determine whether an immune response has been generated. The primary intended use of the large number of serological assays to detect SARS-CoV-2 antibodies (only a few detected neutralizing antibodies) developed and approved under the FDA COVID-19 EUA was for assessing prior SARS-CoV-2 infection and not to determine immune status or whether vaccination was effective. Indeed, recent guidelines published by the Infectious Disease Society of America (IDSA) recommended that: "In individuals with previous SARS-CoV-2 infection or vaccination, the IDSA panel suggests against routine serologic testing given no demonstrated benefit to improving patient outcomes (conditional recommendation, very low certainty of evidence)"29, consistent with CDC recommendations for use of antibody testing: "Antibody tests are not recommended or authorized by the FDA to assess someone’s immunity after COVID-19 vaccination or determine if they need to be vaccinated."5 This may be because no consensus has yet been reached on what neutralizing antibody threshold constitutes a correlate of protection from disease28.

This seems to contrast with the design of clinical trials of SARS-CoV-2 vaccines whose key data include immune responses30, as evidenced in the FDA approval for Moderna’s bivalent vaccine (Original and Omicron BA.1) in 2023 that was based on both safety and immunogenicity data from a clinical trial31. In addition, mathematical models to estimate a correlate of protection threshold based on viral load and neutralizing antibody data have been developed32,33,34, suggesting that it may be feasible to test for protective immune status as is done for other viral vaccines35. It is also clear that vaccine induced immune responses decline 4–sixfold by 6 months following vaccination36, and that neutralizing antibodies are part of a complex immunological response to infection37. Despite these considerations, neutralizing antibodies may be the best, if not flawed, practical and economically feasible test that could provide important information for those who may be at risk for a less-than-optimal response to vaccination such immunocompromised or older individuals. Developing assays to detect neutralizing antibodies elicited by vaccines rather than by natural infection, may avoid the impact of SARS-CoV-2 variants that alter the antigenicity of the S1-RBD binding which may decrease the accuracy of assays based on pre-variant S1-RBD sequences. Neutralizing antibody assays that are based on vaccine S1-RBD sequences would not be affected by SARS-CoV-2 variants. Measuring vaccine response could potentially be used to guide subsequent vaccine boosting38.

There are several limitations to this study. It was conducted using the initial Wuhan strain S1-RBD sequence. Although all Pfizer and Moderna vaccines have included this sequence, the boosters now have incorporated variant strain sequences and will likely do so in the future. We plan to produce a lateral flow device that includes the next vaccine booster sequences. We also chose to use plasma and whole blood drawn by venipuncture rather than finger stick samples. We had previously conducted a study in which several hundred individuals participated in self-collection and performance of a finger-stick lateral flow device and observed tremendous variability in blood flow and ease of sample collection. Nevertheless, we plan to extend the use of the system reported here to finger-stick blood collection.