Skip to main content

Table 5 Zero-inflated regression analysis of factors associated with abdominal aortic calcification, taking Negative binomial distribution to model KI scores

From: A low fractional excretion of Phosphate/Fgf23 ratio is associated with severe abdominal Aortic calcification in stage 3 and 4 kidney disease patients

Model 1     Model 2   
Variable Exp(B) 95% CI p Exp(B) 95% CI p
KI score     KI score   
Intercept 1.568 0.897: 2.740 0.114 2.023 1.090; 3.756 0.026
Male sex 1.086 0.837; 1.410 0.534 1.115 0.830; 1.498 0.471
Age (y) - 50 1.022 1.004; 1.040 0.015 1.010 0.989; 1.032 0.349
CKD Stage 4 vs. 3 1.232 0.941; 1.612 0.128 N/A N/A N/A
Carotid plaque 1.266 0.844; 1.897 0.254 1.577 0.943; 2.636 0.082
Cutoff point       
((FEP/FGF23) < 1/3.9) 1.817 1.422; 2.320 <0.001 1.847 1.374; 2.482 <0.001
Inflation in KI = 0     Inflation in KI = 0   
Intercept 17.629 4.113; 75.551 <0.001 6.329 1.506; 26.596 0.012
Male sex 0.250 0.092; 0.682 0.007 0.305 0.078; 1.189 0.087
Age (y) - 50 0.919 0.875; 0.965 0.001 0.905 0.846; 0.969 0.004
CKD Stage 4 vs. 3 0.347 0.128; 0.942 0.038 N/A N/A N/A
Carotid plaque 0.271 0.100; 0.738 0.011 0.337 0.080; 1.410 0.136
Cutoff point       
((FEP/FGF23) < 1/3.9) 1.047 0.410; 2.677 0.923 0.958 0.273; 3.356 0.946
  1. Exp (B) for the first and second intercepts are the estimated KI score and the estimated odds of KI = 0 respectively for a patient with the reference categories or zero values of the explanatory variables in the model. Additionally Exp (B) of the predictor measures the odds ratio (B) associated to the variable category or 1 unit change depending of the nature of the variable. The table provides with Exp (B) and 95% confidence intervals (CI) for variables with a statistically significant (p < 0.05) contribution to explain the score of abdominal aortic calcification (AAC) measured by KI and the high frequency of no ACC (KI = 0) simultaneously in a multivariate Zero-inflated regression model in all patients (Model 1), or among patients with an estimated GFR below 30 ml/min (Model 2). The ratio FEP/FGF23 was introduced as a binary variable with a cutoff point of (FEP/FGF23) < (1/3.9).