Table 2.

Accuracy of using different combinations vs the single biomarker of hemoglobinuria for the detection of CKD in patients with SCA

Model (biomarker combination)Sensitivity (n = 22), % (n)Specificity (n = 23), % (n)PPV, %NPV, %YI, %AUC
AUC valuePP
A (Hb) 100.0 (22) 69.6 (16) 69.5 100.0 69.6 0.858 Reference 0.018 
B (Hb + CP + ORM) 95.5 (21) 82.6 (19) 79.2 96.3 78.1 0.948 0.018 Reference 
1 (CP) 77.3 (17) 91.3 (21) 86.1 85.3 68.6 0.843 0.938 0.011 
2 (ORM) 90.9 (20) 82.6 (19) 78.4 92.9 73.5 0.868 0.729 0.039 
3 (Hb + CP) 77.3 (17/22) 91.3 (21/23) 86.1 85.3 68.3 0.923 0.059 0.245 
4 (Hb + ORM) 90.9 (20/22) 86.9 (20/23) 82.9 93.2 77.9 0.919 0.069 0.272 
5 (CP + ORM) 95.5 (21/22) 78.3 (18/23) 75.3 96.1 73.8 0.927 0.177 0.214 
Model (biomarker combination)Sensitivity (n = 22), % (n)Specificity (n = 23), % (n)PPV, %NPV, %YI, %AUC
AUC valuePP
A (Hb) 100.0 (22) 69.6 (16) 69.5 100.0 69.6 0.858 Reference 0.018 
B (Hb + CP + ORM) 95.5 (21) 82.6 (19) 79.2 96.3 78.1 0.948 0.018 Reference 
1 (CP) 77.3 (17) 91.3 (21) 86.1 85.3 68.6 0.843 0.938 0.011 
2 (ORM) 90.9 (20) 82.6 (19) 78.4 92.9 73.5 0.868 0.729 0.039 
3 (Hb + CP) 77.3 (17/22) 91.3 (21/23) 86.1 85.3 68.3 0.923 0.059 0.245 
4 (Hb + ORM) 90.9 (20/22) 86.9 (20/23) 82.9 93.2 77.9 0.919 0.069 0.272 
5 (CP + ORM) 95.5 (21/22) 78.3 (18/23) 75.3 96.1 73.8 0.927 0.177 0.214 

Criteria or cutoffs for each model was model A, ≥0.529 ng/mg; model B, ≥2 markers; model 1, ≥1.747 mg/g; model 2, ≥5.750 mg/g; and model 3 to model 5, ≥2 markers.

A test for the equality of the AUC compared with the simple and complete models was performed. Receiving operating characteristic curves were constructed to determine appropriate cutoffs for each biomarker based on the highest values of YI.

PPV and NPV were calculated in the scenario that the prevalence of patients with stage 1 CKD is 48.9% (95% confidence interval, 33.7%-64.2%).

Bold letters were used to indicate the reference model.

NPV, negative predictive value.

or Create an Account

Close Modal
Close Modal