Table 4

Resistin as independent variable in logistic regressions and test variable in area under ROC curves

Logistic regression with resistin as independent variablea

ROC curve with resistin as test variableb


P value

Area under ROC curve (95% CI)


Erythrocyte sedimentation rate

0.001

0.66 (0.58 to 0.75)

S-creatinine

0.19

0.55 (0.46 to 0.65)

Z score hip total

0.2

0.57 (0.46 to 0.67)

Z score radius total

0.13

0.56 (0.46 to 0.66)


aErythrocyte sedimentation rate (ESR) and S-creatinine were defined as normal or pathological according to standard laboratory normal values. These variables were dependent in logistic conditional forward regression analyses with resistin as independent variable. The same method was used for z score hip total and radius total with the cutoff value of -1 standard deviation (SD). bArea under the receiver operating characteristic (ROC) curve was calculated with ESR (normal or not), S-creatinine (normal or not), z score hip total, and z score radius total (above or below -1 SD) as the state variable and resistin as test variable. Null hypothesis: true area under ROC curve = 0.5. CI, confidence interval.

Almehed et al. Arthritis Research & Therapy 2008 10:R15   doi:10.1186/ar2366

Open Data