Skip to main content
. 2021 May 19;59(6):e03161-20. doi: 10.1128/JCM.03161-20

TABLE 5.

Calculated clinical accuracy and kappa statistic for concordance of the commercial methods for the detection of anti-measles IgM antibodies against the predetermined classification of the measles and non-measles sera (n = 52 and 187, respectively)a

Method Measles specimens with equivocal results counted as negativeb
 All specimens with equivocal results counted as positivec
Accuracy (%) 95% CI (%) Kappa statistic 95% CI Accuracy (%) 95% CI (%) Kappa statistic 95% CI
Enzygnost 95.0 91.2–97.3 0.858 0.781–0.936 95.8 92.2–97.9 0.884 0.813–0.954
Euroimmun 92.1 87.7–95.0 0.762 0.660–0.863 94.6 90.7–96.9 0.843 0.761–0.926
Euroimmun Nucleoprotein 93.7 89.7–96.3 0.800 0.704–0.897 96.2 92.7–98.2 0.885 0.812–0.959
IBL 95.8 92.2–97.9 0.880 0.808–0.953 96.2 92.7–98.2 0.893 0.825–0.961
LIAISON XL 96.7e 93.3–98.4 0.897 0.828–0.967 96.7 93.3–98.4 0.897 0.828–0.967
Microimmune 96.2 92.7–98.2 0.890 0.820–0.960 97.5e 94.4–99.0 0.928 0.872–0.985
NovaLisa 90.0 85.3–93.3 0.738 0.641–0.835 90.0e 85.3–93.3 0.738 0.641–0.835
Serion (activity calculator)d 90.8 86.2–94.0 0.763 0.670–0.855 90.8e 86.2–94.0 0.763 0.670–0.855
Serion (OD range)d 91.2 86.7–94.3 0.772 0.681–0.863 91.2e 86.7–94.3 0.772 0.681–0.863
Serion (special case formula)d 89.1e 84.3–92.6 0.726 0.630–0.822 89.1e 84.3–92.6 0.726 0.630–0.822
a

Measles samples with equivocal results were counted as both negative and positive. Non-measles samples with equivocal results were always counted as positive.

b

Specimens in the measles sera panel with equivocal results were counted as negative, and specimens in the non-measles sera panel with equivocal results were counted as positive. In this scenario, equivocal results are considered to be “always wrong.”

c

All specimens (measles and non-measles) with equivocal results were counted as positive. In this scenario, equivocal results for the measles sera panel were considered correct and incorrect for the non-measles sera panel.

d

Three methods of sample result determination, using the single set of optical density data from the test plates, were provided in the manufacturer’s IFU. All three methods were evaluated.

e

Significant difference (P < 0.05) between the most accurate and the least accurate method(s) based on nonoverlapping 95% confidence intervals. When equivocal results were considered to be “always wrong,” the most accurate method, LIAISON XL, was significantly (P < 0.05) better than the Serion kit using the special case formula result determination method. When equivocal results were counted as positive, the most accurate method, Microimmune, was significantly (P < 0.05) better than NovaLisa and Serion (all three result determination methods).