Qualitative Variation among Commercial Immunoassays for Detection of Measles-Specific IgG

Measurement of measles virus-specific IgG is used to assess presumptive evidence of immunity among immunocompetent individuals with uncertain immune or vaccination status. False-negative test results may lead to unnecessary quarantine and exclusion from activities such as employment, education, and travel or result in unnecessary revaccination. In contrast, false-positive results may fail to identify susceptible individuals and promote spread of disease by those who are exposed and unprotected. To better understand the performance characteristics of tests to detect measles IgG, we compared five widely used, commercially available measles IgG test platforms using a set of 223 well-characterized serum samples.

ABSTRACT

Measurement of measles virus-specific IgG is used to assess presumptive evidence of immunity among immunocompetent individuals with uncertain immune or vaccination status. False-negative test results may lead to unnecessary quarantine and exclusion from activities such as employment, education, and travel or result in unnecessary revaccination. In contrast, false-positive results may fail to identify susceptible individuals and promote spread of disease by those who are exposed and unprotected. To better understand the performance characteristics of tests to detect measles IgG, we compared five widely used, commercially available measles IgG test platforms using a set of 223 well-characterized serum samples. Measles virus neutralizing antibodies were also measured by in vitro plaque reduction neutralization, the gold standard method, and compared to IgG test results. Discrepant results were observed for samples in the low-positive ranges of the most sensitive tests, but there was good agreement across platforms for IgG-negative sera and for samples with intermediate to high levels of IgG. False-negative test results occurred in approximately 11% of sera, which had low levels of neutralizing antibody.

INTRODUCTION

Routine two-dose vaccination led to the elimination of endemic measles in the United States in 2000 (1–4). However, in 2019, the United States experienced the highest annual number of measles cases since 1992, largely due to repeated introductions by U.S. residents traveling abroad and returning to undervaccinated communities. Most cases (75%) were associated with two outbreaks in New York Orthodox Jewish communities that threatened the elimination status of measles in the United States. Approximately 89% of measles patients were unvaccinated or had unknown vaccination status, and 11% were vaccine recipients (5). The effectiveness of two-dose MMR vaccination for protection against measles is approximately 97% (range, 67 to 100%) (6). Although some vaccine recipients who are exposed may develop symptomatic infection, they are less likely to transmit virus (7–10) and often have a milder clinical presentation (11–15). Measles virus-specific neutralizing IgG antibody is essential for protection against disease, and a serum titer of >120 is considered protective (16). A serum titer of 120 was later extrapolated to be 200 mIU/ml using the first World Health Organization (WHO) International Serum Standard (IS) and 120 mIU/ml using the second WHO IS (17).

Effective outbreak control relies upon interrupting chains of transmission by identification and vaccination of susceptible individuals and by quarantine of those who are likely to spread disease following exposure. Importantly, measles vaccination can provide effective postexposure prophylaxis (PEP) if given within 72 h of initial exposure (18). Although it may be reasonable to vaccinate most individuals with uncertain measles immune status during an outbreak, vaccination is contraindicated for certain medical conditions (19).

In an outbreak setting, when immune status is uncertain, timely and accurate determination of measles IgG seropositivity can provide important guidance for individual health care decisions. There are a variety of commercial measles IgG test platforms that differ by specific test antigen(s) and method of detection. Accordingly, the performance characteristics among tests differ with respect to sensitivity and specificity. Although relatively infrequent, discrepant IgG test results have been observed among individuals who have received one or more MMR vaccinations. During outbreaks, there is an increase in measles IgG testing volumes due to case-contact investigations and healthy individuals seeking evidence of presumptive immunity. The purpose of this study was to define the performance characteristics of five commercially available measles IgG test platforms that are in common use among commercial and public health laboratories and to compare these results with results from the standard reference assay for determining measles immunity, plaque reduction neutralization (PRN) (16).

MATERIALS AND METHODS

Sera.

A set of 223 previously characterized deidentified sera were chosen for this study based on available volumes and previous measles IgG and IgM test enzyme-linked immunosorbent assay (ELISA) results. The Centers for Disease Control and Prevention (CDC) Human Research Protections Office determined that this study was exempt from Institutional Review Board (IRB) review.

Kit descriptions.

The five measles IgG test platforms that were compared included Vidas measles IgG (bioMérieux, Durham NC; catalog no. 30219), BioPlex 2200 MMRV IgG (Bio-Rad, Hercules, CA; catalog no. 6652450), Liaison measles IgG (Diasorin, Cypress, CA; catalog no. 318810), Captia measles IgG (Trinity Biotech, Jamestown, NY; catalog no. 2326000), and the measles IgG test system (Zeus Scientific, Branchburg, NJ; catalog no. 9Z9271G). All sera were tested by each platform according to the manufacturers’ instructions. The Zeus and Trinity tests are manually performed ELISAs, and the remaining tests are performed with manufacturer-specific automated equipment.

Of note, the Diasorin Liaison measles IgG platform has historically been marketed in the United States and abroad with two different cutoff values for the interpretation of results. Specifically, for kits sold in the United States, the interpretation of results has been as follows: <25.0 arbitrary units (AU)/ml is IgG negative, ≥25.0 AU/ml and <30.0 AU/ml is equivocal, and ≥30.0 AU/ml is positive. However, outside the United States, the interpretation of results has been as follows: <13.5 AU/ml is IgG negative, ≥13.5 to <16.5 AU/ml is equivocal, and ≥16.5 AU/ml is positive. Recently, the manufacturer lowered the interpretation of results in the United States to match the interpretation given for kits sold outside of the United States. For this study, results of the DiaSorin Liaison measles IgG assay were interpreted with both sets of criteria. As described below, this change in cutoff has a minor overall effect on the sensitivity and specificity of the test but could make important differences for certain individuals with low levels of IgG.

Measles in vitro plaque reduction neutralization.

Serum neutralizing antibody titers were measured by PRN as previously described (17). In this study, the WHO second IS anti-measles serum (IS, coded 66/202, supplied by the National Institute for Biological Standards and Control, South Mimms, United Kingdom) was included to calculate the reciprocal of the 50% endpoint titer by the Kärber method (20). A dilution of the WHO second IS was included in each assay, and endpoint titers for the internal standards did not differ by more than 20%. With the validation of the second WHO standard serum, PRN titers were expressed in mIU/ml. A titer of 8 was equivalent to a concentration of 8 mIU/ml. PRN titers were used as the gold standard for comparison of the various IgG test platforms. Receiver-operator curves were determined for each IgG test platform using PRN titers of 8, 40, 60, and 120 mIU/ml as references. A PRN titer of >120 was considered protective for measles (16) and is equivalent to 120 mIU/ml based on use of the WHO second international standard.

Measles IgM capture ELISA.

Sera were tested for measles IgM using a capture ELISA as previously described (21). Because measles-specific IgM antibody can neutralize virus and contribute to the PRN titer, sera that were IgM positive were analyzed as a separate group in the platform comparison as described below.

RESULTS

A set of 223 sera was examined by measles in vitro PRN, measles-specific IgM capture ELISA, and five different commercially available measles IgG tests. The methods and antigens used by the various IgG test platforms may or may not detect neutralizing antibody. However, we chose to directly compare each test to PRN titer since >120 mIU/ml is considered protective (16). In addition, because it is not practical to perform PRN for routine testing, IgG test methods that best correlate with PRN would be the most useful in determining measles immunity.

Sera containing high levels of measles-specific IgM can contribute detectable neutralizing activity to the PRN assay. To avoid confounding results, ELISA-to-PRN comparisons were performed using only IgM-negative specimens (n = 146). We performed receiver-operating characteristic (ROC) analysis for each IgG test using the PRN results as the reference standard. Inherent variability in the PRN assay can be 2- to 3-fold between each test run for individual sera (17). Previous reports used >8 mIU/ml as a seropositive PRN result and >120 mIU/ml as a seroprotective result. We used ≥8-, ≥40-, ≥60-, and ≥120-mIU/ml PRN cutoffs to determine the effects on sensitivity and specificity for each comparison (Table 1). A PRN titer of ≥40 mIU/ml gave the optimal cutoff for overall agreement, sensitivity, and specificity for all tests. Although these IgG immunoassay tests are not intended to specifically detect neutralizing antibody, an in vitro plaque neutralization titer of 40 mIU/ml generally discriminates between sera that are IgG negative versus positive based on these data. The results of the ROC analysis are summarized in Table 1. Compared to PRN, the Zeus test had the greatest overall agreement (92.5%), sensitivity (91.7%), and specificity (100%). At a PRN cutoff of ≥40 mIU/ml, the remaining commercial assays had moderate overall agreement with PRN (82.2 to 87.7%). Although all the other tests had excellent specificities ranging from 96.0 to 100%, the sensitivity was lower and ranged from 78.3 to 88.2%. Finally, the results from each test platform for the entire set of 223 sera, including 77 measles IgM-positive sera, were compared against the Zeus assay (see Table 1). The overall agreement of the other platforms with the Zeus test ranged from 88.3 to 92.4%. The ranges for sensitivity and specificity were 87.1 to 93.3 and 94.4 to 100% respectively. Except for PRN, there did not appear to be any effect of measles-specific IgM on the performance of these commercial IgG test platforms relative to each other.

TABLE 1.

Performance of PRN versus IgG immunoassays using multiple PRN cutoffs and comparison of Zeus to other IgG immunoassays

Parametera	Assay resultsb
	Zeus					Trinity
	vs Zeus	vs PRN ≥8	vs PRN ≥40	vs PRN ≥60	vs PRN ≥120	vs Zeus	vs PRN ≥8	vs PRN ≥40	vs PRN ≥60	vs PRN ≥120
Total no.		146	146	146	146	223	146	146	146	146
No. agreed		131	135	133	126	206	124	128	129	127
No. discrepant		15	11	13	20	17	22	18	17	19
% discrepant		10.3	7.5	8.9	13.7	7.6	15.1	12.3	11.6	13.0
% agreement		89.7	92.5	91.1	86.3	92.4	84.9	87.7	88.4	87.0
% sensitivity (95% CI)		88.1 (81.1–93.2)	91.74 (85.3–95.9)	92.37 (86.0–96.5)	96.08 (90.3–98.9)	93.3 (88.3–96.6)	38.89 (30.3–48.0)	88.18 (80.6–93.6)	87.3 (79.9–92.7)	93.14 (86.4–97.2)
% specificity (95% CI)		100 (83.2–100.0)	100 (86.3–100.0)	92.86 (76.5–99.1)	70.45 (54.8–83.2)	98.2 (90.1–100.0)	100 (83.2–100.0)	100 (86.3–100.0)	96.43 (81.7–99.9)	79.55 (64.7–90.2)
AUC (95% CI)		0.99 (0.977–1.003)	0.99 (0.987–1.003)	0.98 (0.962–1.001)	0.95 (0.910–0.983)	0.98 (0.950–1.013)	0.99 (0.975–1.006)	0.99 (0.981–1.003)	0.97 (0.948–1.001)	0.92 (0.874–0.970)

^aAUC, area under the curve. AUC provides a measure of separability and predicts the ability of model to distinguish between classes.

^bSee Materials and Methods for information regarding the Diasorin Liaison measles IgG test cutoffs.

Figure 1 shows the comparison between PRN titer and IgG results for 223 individual sera across the 5 IgG platforms. As mentioned previously, because some measles-specific IgM antibodies can have neutralizing activity, sera that were IgM negative (n = 146; Fig. 1A, C, E, G, and I) were analyzed separately from IgM-positive sera (n = 77; Fig. 1B, D, F, H, and J). Most of the commercial assays evaluated demonstrated good correlation with PRN in the lowest ranges (true negative) and highest ranges (strong positive) of the assays. The majority of discrepant results occurred with sera that yielded low-positive PRN results (>8 and <120 mIU/ml) and were typically discrepant across multiple platforms. Individual results highlighted in red in Fig. 1A and B indicate samples that were discrepant between the Zeus test and at least one other IgG platform. Results highlighted in red in Fig. 1C to J were discrepant between the indicated test and the Zeus IgG result.

FIG 1.

Comparison of quantitative IgG test results with PRN. Measles IgM-negative samples are shown in panels A, C, E, G, and I; measles IgM-positive samples are shown in panels B, D, F, H, and J. The Zeus test had the greatest overall agreement with PRN and was therefore used as a comparator. The platforms tested include Zeus (A and B), Trinity (C and D), Bio-Rad (E and F), bioMérieux Vidas (G and H), and Diasorin (I and J). Horizontal dotted lines indicate cutoffs provided by the test manufacturers. For Diasorin (I and J), the black horizontal lines indicate the cutoff previously used in the United States, and the blue horizontal lines indicate the new cutoff used in the United States and also used in Europe. Vertical dotted lines indicate PRN = 40 and 120 mIU/ml. Samples indicated in red in panels A and B were discrepant between Zeus and at least one other platform. Samples indicated in red in panels C to J were discrepant with the Zeus result.

Thirty samples with sufficient volume were chosen for reproducibility testing across all test platforms (data not shown). Qualitative results were compared and none of the platforms had more than two discrepant results (range, 0 to 6.7%). All discrepant results were near the cutoffs for the negative-indeterminate or indeterminate-positive ranges.

DISCUSSION

We compared five measles IgG test platforms in common use among commercial and public health laboratories in the United States; IgG results were compared to neutralizing antibody titers, taking into consideration the presence of measles-specific IgM and the effects of IgM on virus neutralization. Although none of the IgG tests are designed to specifically measure neutralizing antibody, all of them demonstrated a good overall correlation with PRN (Fig. 1). Furthermore, measles IgM did not appear to cause interference with any of the IgG tests that were examined, since a wide range of IgG values were detected among IgM-positive samples across all test platforms.

The results demonstrate differences in the sensitivity and specificity of individual IgG tests. In general, tests that are based on a manual ELISA format (Zeus and Trinity) were more sensitive than the automated high-throughput platforms (BioPlex 2200, Diasorin Liaison, and bioMérieux Vidas). However, all the commercial platforms demonstrated good agreement of qualitative results. Similar to previous reports, up to 11% of samples gave discordant results in comparisons of the most-sensitive versus the least-sensitive platforms (22, 23). The discrepant results were in the low-positive, equivocal, and high-negative ranges for all platforms.

In an outbreak setting, there is often an increased level of IgG testing. Since most commercial assay cutoffs are set to identify true negatives, increased testing will result in an increase in the total number of potentially false-negative results. Previously vaccinated individuals with low levels of IgG may have false-negative results depending on the test that is used. This may lead to either unnecessary vaccination, additional laboratory testing or unnecessary restriction of activities as well as a waste of public health resources.

The PRN test is considered the gold standard laboratory method for determining measles immunity because it measures functional neutralizing antibody. From a public health perspective, it is not practical or necessary to determine PRN titers for every individual that requires testing (16). Compared to IgG testing by ELISA, PRN is labor-intensive, time-consuming, and costly, and it has a long turnaround time. PRN, therefore, is not practical for routine testing.

In a measles elimination setting, a conservative approach for an IgG test would be to have an appropriately high cutoff that may err on the side of providing a small fraction of false-negative results. This approach would allow the identification of individuals who may not be sufficiently protected and who should consider (re)vaccination. In contrast, an IgG test with a cutoff that is too low would result in false-positive results, which would prevent the identification and vaccination of susceptible individuals. Additional testing guidance is available from the CDC (https://www.cdc.gov/measles/hcp/index.html#lab) and the American Public Health Laboratories (https://www.aphl.org/programs/infectious_disease/_layouts/15/WopiFrame.aspx?sourcedoc=/programs/infectious_disease/Documents/APHL%20statement%20on%20measles%20IgG%20final.docx&action=default&DefaultItemOpen=1). The data presented here are intended to raise awareness regarding the limitations and utility of measles IgG testing in general and to provide an estimate of the performance characteristics of the specific tests described. Individuals that have negative or equivocal results for measles IgG should be vaccinated or revaccinated (6). In some cases, vaccination is not possible and testing with a second line diagnostic assay may be necessary to determine immune status.