We evaluated the effect of elevating the IRT-cut-off from 50 to 65 渭g/l (~ 97.5th to ~ 99.0th percentile), the need of a failsafe protocol (FS, IRT 鈮?#xA0;99.9th percentile) and the relative performance using either two IRT-dependent PAP-cut-offs or one PAP-cut-off.
Elevation of the IRT cut-off to 65 渭g/l (~ 99.0th percentile) increased the PPV significantly (Dresden: 0.065 vs. 0.080, p < 0.0001, Prague: 0.052 vs. 0.074, p < 0.0001) without reducing sensitivity. All three IRT/PAP protocols showed a trend towards a higher sensitivity with FS than without and when using one PAP-cut-off instead of two IRT-dependent PAP-cut-offs.
For best performance we suggest an IRT/PAP protocol with an IRT-cut-off close to the 99.0th percentile, FS, and a single PAP-cut-off.