Association of dietary phosphorus intake and phosphorus to protein ratio with mortality in hemodialysis patients

Abstract

Background and objectives: Epidemiologic studies show an association between higher predialysis serum phosphorus and increased death risk in maintenance hemodialysis (MHD) patients. The hypothesis that higher dietary phosphorus intake and higher phosphorus content per gram of dietary protein intake are each associated with increased mortality in MHD patients was examined.

Design, setting, participants, & measurements: Food frequency questionnaires were used to conduct a cohort study to examine the survival predictability of dietary phosphorus and the ratio of phosphorus to protein intake. At the start of the cohort, Cox proportional hazard regression was used in 224 MHD patients, who were followed for up to 5 years (2001 to 2006).

Results: Both higher dietary phosphorus intake and a higher dietary phosphorus to protein ratio were associated with significantly increased death hazard ratios (HR) in the unadjusted models and after incremental adjustments for case-mix, diet, serum phosphorus, malnutrition-inflammation complex syndrome, and inflammatory markers. The HR of the highest (compared with lowest) dietary phosphorus intake tertile in the fully adjusted model was 2.37. Across categories of dietary phosphorus to protein ratios of <12, 12 to <14, 14 to <16, and > or =16 mg/g, death HRs were 1.13, 1.00 (reference value), 1.80, and 1.99, respectively. Cubic spline models of the survival analyses showed similar incremental associations.

Conclusions: Higher dietary phosphorus intake and higher dietary phosphorus to protein ratios are each associated with increased death risk in MHD patients, even after adjustments for serum phosphorus, phosphate binders and their types, and dietary protein, energy, and potassium intakes.

Figure 1.

Scatter plots (including the regression line and 95% CI) reflecting the correlations of dietary phosphorus intake with dietary protein intake (upper panel) and serum phosphorus concentration (lower panel).

Figure 2.

Cubic spline models of the Cox proportional regression analyses reflecting adjusted mortality predictability (with 95% CI) according to the percentile of the patient’s dietary phosphorus intake in the entire cohort of 224 MHD patients over 5 years (from October 2001 to January 2007). Spline models are with two degrees of freedom. Case-mix variables include age, gender, race/ethnicity, diabetes mellitus, dialysis vintage, insurance, marital status, modified Charlson comorbidity score, dialysis dose (Kt/V), intake of phosphorus binders, and residual urine. Dietary variables include energy, protein, and potassium intake. MICS variables include serum concentrations of albumin, creatinine, bicarbonate, ferritin, calcium, and phosphorus, blood levels of hemoglobin, white blood cell, and lymphocyte percent; and nPCR, body mass index, and averaged doses of erythropoietin and injected active vitamin D. Inflammatory markers include serum concentrations of CRP, IL-6, and TNF-α.

Figure 3.

Cubic spline models of the Cox proportional regression analyses reflecting adjusted mortality predictability (with 95% CI) according to the percentile of the patient’s dietary phosphorus to protein ratio in the entire cohort of 224 MHD patients over 5 years (from October 2001 to January 2007). Spline models are with two degrees of freedom. Case-mix variables include age, gender, race/ethnicity, diabetes mellitus, dialysis vintage, insurance, marital status, modified Charlson comorbidity score, dialysis dose (Kt/V), intake of phosphorus binders, and residual urine. Dietary variables include energy, protein, and potassium intake. MICS variables include serum concentrations of albumin, creatinine, bicarbonate, ferritin, calcium, and phosphorus, blood levels of hemoglobin, white blood cell, and lymphocyte percent; and nPCR, body mass index, and averaged doses of erythropoietin and injected active vitamin D. Inflammatory markers include serum concentrations of CRP, IL-6, and TNF-α.