P1394ASSOCIATION BETWEEN PARATHYROID HORMONE AND MORTALITY IN HAEMODIALYSIS: THE DIABETES MAKES THE DIFFERENCE

Abstract

Abstract Background and Aims Diabetes is the most common cause of chronic kidney disease (CKD) stage 5 and a relevant risk factor for mortality. The available CKD-MBD guidelines do not make differences between diabetics and non-diabetics regarding the optimal targets of the serum bone and mineral biochemical parameters. Thus, the objective of this analysis of COSMOS was to compare the association between the main serum bone and mineral biochemical parameters (calcium-Ca, phosphorus-P and parathyroid hormone-PTH) and mortality in diabetic and non-diabetic patients in haemodialysis. Method COSMOS is a 3-year, observational, prospective, open-cohort study that includes 6797 patients from 227 dialysis centres that were randomly selected from 20 European countries. Cox proportional hazard regression models with penalized splines smoothing were used to analyse the association between relative risk of mortality vs. Ca, P or PTH. Hazard ratios were adjusted by demographics, comorbidities, treatments and biochemical parameters. Interaction analysis was used to analyse if diabetes can influence the association between relative risk of mortality and Ca, P and PTH. Results The present analysis includes 6306 patients with follow-up data and accurate information on diabetes status. The mortality rate in COSMOS was 13.3 deaths per 100 patient-years (17.8 in diabetics and 11.4 in non-diabetics). At baseline, diabetic patients were older (66.7 ± 11.7 vs. 62.8 ± 15.3 years, p <0.001), had a higher body mass index (27.3 ± 5.6 vs. 24.5 ± 4.5 kg/m2, p <0.001) and more frequent history of cardiovascular disease (83.1% vs. 67.2%, p<0.001). The percentage of patients prescribed drugs to correct the bone and mineral biochemical parameters such as, phosphate binding agents, vitamin D receptor activators and calcimimetics was lower in diabetics (82.3% vs. 86.4%, 45.4% vs. 48.5% and 4.6% vs. 6.9% respectively, p < 0.03). There were no significant differences in serum P (5.3 ± 1.3 vs. 5.4 mg/dL ± 1.5, p = 0.3) and haemoglobin (11.4 ± 1.4 vs. 11.4 ± 1.4 g/dL, p = 0.5) but diabetic patients showed significant -but not clinically relevant- lower levels of Ca (9.0 ± 0.7 vs. 9.1 ± 0.8 mg/dL, p <0.001), PTH (median [interquartile range]: 199 [109, 335] vs. 216 [108, 396] pg/mL, p = 0.001) and albumin (3.7 ± 0.5 vs. 3.8 ± 0.5 g/dL, p <0.001). No significant interaction was found between diabetes and serum Ca or P and their association with relative risk of mortality. However, diabetes and PTH showed a marginal significant interaction in their association with mortality (p=0.050). The slopes of the curve of relative risk of mortality vs. PTH were more pronounced in diabetics (see figure). High serum PTH was significantly associated with a higher relative risk of mortality in diabetics but not in non-diabetics. The range of serum PTH where the minimum risk of mortality was observed (defined as serum PTH with a HR<1.1, i.e. less than 10% increase in the relative risk of mortality) was narrower in the diabetics compared with non-diabetics (204-546 vs. 166-757 pg/mL, figure) Conclusion These results suggest for the first time a differential effect of diabetes on the association between relative risk of mortality and serum PTH. High serum PTH predicts death only in diabetics. These findings could have relevant implications for the diagnosis and treatment of CKD-MBD in diabetic and non-diabetic patients in haemodialysis.