MO499CKD-MBD BIOMARKERS AND CKD PROGRESSION: AN ANALYSIS BY JOINT MODELS

Abstract

Abstract Background and Aims Biomarkers of the Chronic Kidney Disease – Bone Mineral Disorder (CKD-BMD)-including serum phosphate, calcium, PTH, 1,25 vitamin D, alkaline phosphatase and FGF23- have been implicated in CKD progression in follow up studies focusing on single measurements made at baseline only. The relationship between the time trend of these biomarkers over the course of CKD and renal outcomes has never been tested. Method In this study we applied the joint model, an approach combining the mixed linear model (MLM) and Cox’s regression analysis, to investigate the longitudinal relationship between repeated measurements (median 3, interquartile range 2-5) of serum phosphate, calcium, PTH and alkaline phosphatase and a composite renal outcome measure (eGFR reduction >30%, dialysis or transplantation) cohort of 729 stage 2-5 CKD patients (average eGFR = 36±13 ml/min/1.73 m2) over a 2.4 year follow up. Data were adjusted for traditional and CKD specific risk factors in models including also baseline serum 1,25 vitamin D and FGF23. Results On univariate MLM analyses, serum PTH was the sole CKD-BMD biomarker which significantly changed over time [b=0.030 (95% CI from 0.018 to 0.041), p<0.001]. In the multivariate mixed linear sub-model of the Joint Model, repeated measurements of PTH were related directly with male gender (P=0.033), alkaline phosphatase (P<0.01) and systolic BP (P=0.004) and inversely with calcium (P<0.001), baseline eGFR (P<0.001), and CRP (P<0.045). In the survival analysis sub-model of the joint model, the longitudinal series of PTH valuesover time were directly related to the incidence rate of renal events [HR (1 ln increase in PTH)=3.358 (95% CI from 2.628 to 4.293), p<0.001] and this was also true after data adjustment for a series of potential confounders [HR (1 ln increase in PTH)=2.25 (95% CI from 1.524 to 3.321), p<0.001]. In this analysis, also serum phosphate, baseline FGF23, proteinuria, baseline 1.25 vit D, calcium, and haemoglobin were significantly related to the incidence rate of renal events and these relationships were all independent of the baseline eGFR. Conclusion In this longitudinal study in stage 2-5 CKD patients investigating the links of CKD-MBD biomarkers with a composite renal outcome, PTH was the sole variable to show a linear increase overtime and to be directly related with the same outcome. Even though stable over-time, also the longitudinal series of serum phosphate (in a direct fashion) and calcium (in an inverse fashion) values were related to the same outcome as it were baseline FGF23 and 1.25 vit D. Findings in this longitudinal study further underscore the potential relevance of CKD-MBD biomarkers as risk factors for CKD progression.