The impact of high serum phosphorus in the general population is still debated. Studies are heterogeneous, most lack an adjustment for parathyroid hormone, vitamin D and phosphorus intake and the effect might differ by gender and renal function. We investigated the association between serum phosphorus and mortality in American adults.
We prospectively analyzed 5698 non-pregnant and non-CKD adults from the National Health and Nutrition Examination Survey (NHANES) 2003-2006. Serum phosphorus and potential confounders including parathyroid hormone, 25(OH)vitamin D and phosphorus intake were evaluated. All-cause, cardiovascular- and cancer-related deaths were recorded through December 31st, 2015. Sex-specific terciles of serum phosphorus were used to fit adjusted Cox proportional hazard models for mortality. Analysis was stratified by gender and renal function.
A total of 590 deaths were recorded over a median follow-up of 81 months. Women showed higher serum phosphorus than men. The adjusted hazard ratio (HR) for all-cause mortality was 1.35 (95% CI 1.08-1.58) (p = 0.033) for the third tercile (versus second tercile). This increased risk was present in participants with estimated glomerular filtration rate (eGFR) below 90 ml/min/1.73 m but not above, although interaction was not significant (p = 0.12). Interaction by gender, phosphorus intake, PTH and fasting time was also not detected. For cardiovascular and cancer mortality, the adjusted HR was 0.81 (95% CI 0.33-2.00) (p = NS) and 1.45 (95% CI 0.77-2.72) (p = NS), respectively.
We demonstrated that the highest tercile of serum phosphorus is associated with increased all-cause mortality, irrespective of PTH, 25(OH)vitamin D or phosphorus intake. This association may differ by gender and renal function, but larger studies testing for effect modification are needed.

Author