Skip to main content
. 2009 Sep 23;16(11):1576–1582. doi: 10.1128/CVI.00205-09

TABLE 5.

Statistical performance of serology assays using cutoff values recommended by the manufacturer or determined by logistic regression analysisa

Parameter Value(s) for indicated assay
EMA (titer) tTG-IgA Celikey (U/ml) tTG-IgA Immulite 2000 (U/ml) DGP IgA+IgG screen (U) tTG-IgG Celikey (U/ml)
Cutoff points recommended by the manufacturer
    Cutoff value <1/5 (−), ≥1/5 (+) <5 (−), 5-8 (BL), >8 (+) <4 (−), ≥4 (+) <20 (−), 20-30 (w+), > 30 (s+) <7 (−), 7-10 (BL), >10 (+)
    Sensitivity (%) 95.3 94.1 92.9 95.3 67.7
    Specificity (%) 74.2 87.1 74.2 64.5 93.5
    PPV (%) 91.0 95.2 90.8 83.3 96.5
    NPV (%) 85.2 84.4 79.3 88 49.2
    Test efficiency (%) 89.6 91.4 87.9 87 72.4
Cutoff points determined by logic regression analysis
    Cutoff value <1/10 (−), ≥1/10 (+) <7 (−), ≥7 (+) <10 (−), ≥10 (+) <34 (−), ≥34 (+) <3.6 (−), ≥3.6 (+)
    Sensitivity (%) 90.6 90.6 90.6 89.4 81.2
    Specificity (%) 96.8 96.8 87.1 83.9 83.9
    PPV (%) 98.7 98.7 95.1 93.8 93.2
    NPV (%) 78.9 78.9 77.1 74.3 61.9
    Test efficiency (%) 92.2 92.2 89.7 87.9 81.9
a

−, negative; +, positive; w+, weak positive; s+, strong positive; BL, borderline. Borderline and weak-positive results were considered positive for calculating sensitivity, specificity, PPV, NPV, and test efficiency.