Albertine Donker

Chapter 8 288 (95% CI, 87-100%) and the same specificity as in the total study population. The AUC in the mono-allelic affected group versus the IDA group was 0.982 ( p =0.000) with a sensitivity of 90% (95% CI, 72-98%) and a comparable specificity of 100% (95% CI, 93-100%). There was one IRIDA patient with a TSAT/hepcidin ratio of >5.9%, concerning a monoallelic affected patients with a ratio of 9.1%nM, based on a TSAT of 11.8% and a hepcidin level of 1.3 nM after parenteral iron treatment (i.e. intramuscular iron in combination with vitamin C) had been given. Since our choice to imputate hepcidin levels below the lower limit of detection (LLOD) with a value of 0.25 nM might over-estimate the differences between the TSAT/hepcidin ratios of the IRIDA group and IDA control group, we also performed ROC-analysis of the ratio with imputated hepcidin levels of 0.49 nM. Using the latter level for imputation for levels < LLOD, the TSAT/hepcidin ratio also discriminates between IRIDA and IDA with high sensitivity and specificity ( Supplemental Table 1 ). Figure 2. ROC curve of TSAT/hepcidin ratio in IRIDA patients and IDA controls. Receiver operator characteristic curve analysis comparing the TSAT/hepcidin ratio in the non-IRIDA group versus the total IRIDA group (red), the bi-allelic IRIDA group (dark blue) and the mono-allelic IRIDA group (light blue). The arrow indicates the cut-off point of 5.9%/nM, which renders a specificity of 100% and a sensitivity of 95%.