MO938: Variability of Serum Phosphate in Incident Hemodialysis Patients: Association With All-Cause Mortality

Abstract

Abstract BACKGROUND AND AIMS In maintenance hemodialysis, both low and high serum phosphate levels are related to patient outcomes. We aimed to further explore the relation between serum phosphate (P) variability and all-cause mortality. METHOD All adult incident hemodialysis patients treated in U.S. Fresenius Kidney Care clinics between January 2010 and October 2018 were included. The first 6 months on hemodialysis were defined as baseline, months 7–18 as the follow-up period. All-cause mortality was recorded during follow-up. We quantitated variability as directional range (DR), a metric that describes during baseline the difference between the largest and the smallest value; the DR is positive when the smallest value precedes the largest; otherwise, the DR is negative. Cox proportional hazards models with spline terms were applied to explore the association between DR and SD of serum P and all-cause mortality. Additionally, tensor product smoothing splines were computed to study the interactions of P with the DR and SD of P and their joint associations with outcomes. All analyses were adjusted for demographics, comorbidities and nutritional parameters. RESULTS A total of 281 356 patients were included, 258 741 patients were studied. The average age was 62.3 years, 58% were males, 68% were diabetic, 10% with COPD and 25% with CHF. Baseline mean serum P was 5.07 mg/dL, the median DR was + 1.7 mg/dL and the median SD was 0.97. Our results indicate that both higher and lower levels of DR are associated with a higher risk of mortality (Fig. 1). With higher levels of serum P, both lower (negative) and higher levels of DR of serum P were associated with a higher risk of mortality (Fig. 2). In patients with lower levels of serum P, a negative level of DR is associated with a high mortality risk, while the association is attenuated in patients with higher (positive) levels of DR (Fig. 2). On all levels of serum P, higher SD is associated with a higher risk of mortality. CONCLUSION Stable serum P levels between 4 and 6 mg/dL are associated with the best outcomes. Negative DR of serum P is associated with particularly high HR for all-cause mortality in the presence of serum P levels <3 mg/dL, possibly because patients are generally malnourished or inflamed, and a further reduction of serum P indicates nutritional deterioration. Serum P > 8 mg/dL is associated with increased all-cause mortality, irrespective of DR.