MO903: Associations of Serum Sodium With Mortality in Patients on Maintenance Haemodialysis Treatment

Abstract

Abstract BACKGROUND AND AIMS Chronic HD-patients have a 50% 5-year survival and unphysiological fluid and sodium shifts are key pathophysiological contributors. Several studies have shown that hyponatremia is associated with worse outcomes in haemodialysis patients. However, the association of all-cause mortality with volume status in hyponatremic patients is uncertain. Body composition monitoring has provided a more precise tool to estimate fluid status in dialysis patients. The primary aim of this international observational study performed in a large patient database (EUCLID5) was to quantify the association between salt exposure, fluid overload (FO) and mortality risk in haemodialysis patients. METHOD We included 71 706 incident haemodialysis patients from 1 January 2010 to 4 December 2020 in an extended Cox regression model for a time to death analysis. Exposure variables were defined as predialysis serum sodium and fluid overload (in liters as determined by the BCM—Body Composition Monitor using bioimpedance spectroscopy) on a monthly time grid. Effect modifiers, mediators, confounders and predictors were defined a priori. Adults were included with at least one valid BCM measurement within 3 months from first haemodialysis treatment and followed up until death or administrative censoring. All patients included in the study cohort were treated in the NephroCare–FMC dialysis centre network operating in 26 countries. Data were retrieved from the central EuCliD5 database. RESULTS In hyponatremic patients (serum sodium < 135 mmol/L), the association to the risk of death doubled [(HR = 1.97 (1.82–2.12)] in overhydrated patients (relative FO defined as 13% in men, and 15% in women >90th percentile of physiological hydration), was increased by 56% [HR = 1.56 (1.27–1.93)] in underhydrated patients (<–7% <10th percentile of normohydration) and was modestly higher in states of isovolemia (–7%–13% in women and 15% in men). The associated risk of death in states of overhydration decreased [HR = 1.50 (1.27–-1.76)] when patients were hypernatremic (serum sodium >142 mmol/L), (Table 1). Generally, the risk increased with rising FO and declining sodium concentration. CONCLUSION The observed associations to the risk of death reflecting the interplay between states of hypo-, iso- and hypernatremia and states of under, normal and over hydration deserve further study to understand the underlying complex mechanisms and identify therapeutic approaches. When assessing the risk level of different sodium concentrations, fluid status should be taken into account.