Phosphorus binders and survival on hemodialysis

Abstract

Although hyperphosphatemia is a risk factor for mortality, there are limited data on whether therapy with phosphorus binders affects survival. We analyzed a prospective cohort study of 10,044 incident hemodialysis patients using Cox proportional hazards analyses to compare 1-yr all-cause mortality among patients who were or were not treated with phosphorus binders. We performed intention-to-treat analyses to compare patients who began treatment with phosphorus binders during the first 90 d after initiating hemodialysis (n = 3555) with those who remained untreated during that period (n = 5055). We also performed as-treated analyses that modeled phosphorus binder treatment as a time-dependent exposure. We compared survival in a subcohort of treated (n = 3186) and untreated (n = 3186) patients matched by their baseline serum phosphate levels and propensity score of receiving phosphorus binders during the first 90 d. One-year mortality was 191 deaths/1000 patient-years at risk. Treatment with phosphorus binders was independently associated with decreased mortality compared with no treatment in the intention-to-treat, as-treated, and matched analyses. The results were independent of baseline and follow-up serum phosphate levels and persisted in analyses that excluded deaths during the first 90 d of hemodialysis. In summary, treatment with phosphorus binders is independently associated with improved survival among incident hemodialysis patients. Although confirmatory studies are needed in the dialysis setting, future placebo-controlled, randomized trials of phosphorus binders might focus on predialysis patients with chronic kidney disease and normal serum phosphate levels.

Figure 1.

Stratified HRs for mortality in the unmatched subcohort (n = 8610) comparing patients who began treatment with phosphorus binders during the first 90 d on hemodialysis versus those who were not treated during the first 90 d. Models within individual strata were adjusted for all other covariates, including age; gender; race; ethnicity; cause of renal failure; systolic BP; body mass index; standardized mortality rate; initial vascular access; coronary artery disease; congestive heart failure; laboratory variables including albumin, creatinine, calcium, phosphate, PTH, and hemoglobin; and vitamin D therapy at baseline. Percentages represent the proportion of deaths within each stratum, squares represent the HR point estimates, and horizontal lines represent 95% CIs. Point estimates to the left of unity (vertical line) favor a survival benefit of phosphorus binders.

Figure 2.

(A through E) Survival of treated and untreated patients in the overall propensity score–matched cohort (A) and according to quartiles of baseline serum phosphate: <3.7 mg/dl (B), 3.7 to 4.5 mg/dl (C), 4.6 to 5.5 mg/dl (D), and ≥5.6 mg/dl (E).