The principal finding of this study was that serum phosphate was associated with high HD-unit BP, vasoconstriction, and markers of ECD. Phosphate was independently associated with higher pre-HD diastolic BP, TPRI, and ET-1 while controlling for numerous demographic variables. ADMA was associated with both phosphate and PTH in univariate analyses, but there appear to be other confounding variables behind this relationship. Overall, these findings demonstrate a novel relationship between MBD and cardiovascular disease in ESRD patients that may be independent of vascular calcification.
Serum phosphate is associated with increased morbidity and mortality in CKD patients[1, 9], though it is unknown whether BP is directly involved in this relationship. Intervention studies in both animals and healthy humans have demonstrated increases in BP following extended periods of high dietary phosphate intake[10, 11]. One observational study in pre-ESRD CKD patients showed a moderate correlation between serum phosphate and systolic BP which was predominantly found in the patients with diabetes (n = 30), but there was not adjustment for any other variables[12]. Another study in ESRD patients found that diastolic, but not systolic, BP (both measured pre-HD) was higher in patients in the highest tertile of serum phosphate[13]. Our study is novel in that we controlled for numerous demographic variables that have been associated with high pre-HD BP, we analyzed phosphate as a continuous variable, and we included analyses of vasoconstriction measurements and markers of ECD[8, 14].
Some of the proposed mechanisms to explain a relationship between phosphate and BP include increased arterial stiffness, increased renin-angiotensin-aldosterone system (RAAS) activity, and increased sympathetic nervous system (SNS) activity [10, 11, 15, 16]. The negative consequences of increased arterial stiffness include increased central aortic BP (which can contribute to left ventricular hypertrophy) and widened pulse pressure (which can contribute to coronary hypoperfusion.) Unfortunately, the structural changes from vascular calcification may persist even after serum phosphate is acutely brought under control. Our findings of the particularly strong association between phosphate and diastolic BP seem to deviate from the phenotype of isolated systolic hypertension that is often seen with increased arterial stiffness, which suggests that another mechanism may be responsible. As this was a post-hoc analysis from a previously conducted study, we did not have any assessment of RAAS or SNS activity to further evaluate possible explanations for relationship between phosphate and vasoconstriction. However, a large percentage of patients were receiving RAAS inhibiting drugs and/or beta adrenergic receptor antagonists.
Another mechanism proposed to explain the relationship between phosphate and cardiovascular disease is ECD. In an in vitro study, rat aortic ring cells exposed to a high phosphate medium showed significant decrease in dilation compared to those exposed to lower phosphate-containing medium[2]. The same investigators found that healthy humans ingesting a high phosphate meal experienced acute reduction in flow mediated vasodilation post-prandially which was inversely correlated with the serum phosphate level[2]. This study also reported an increase in PTH following phosphate ingestion, which was not accounted for in the analysis. In a community based population study, where only 7% had estimated glomerular filtration rates < 60 mL/min/1.73m2, the investigators found an association between serum phosphate and microvascular dysfunction assessed with skin capillaroscopy[17]. Surprisingly, a cross sectional study in patients undergoing hemodialysis fistula placement found that a U-shaped curve defined the relationship between serum phosphate and FMD independent of BP, race, or presence of diabetes[18]. Parathyroid hormone and other variables related to MBD and nutrition were not taken into consideration, and roughly one third of patients in that study had pre-dialysis CKD. Of note, one clinical trial found that suppression of PTH with intravenous vitamin D analogues improved FMD in CKD patients, but this effect was blunted among patients with persistently high phosphate levels[19]. Collectively, these demonstrate that phosphate and/or additional MBD factors adversely influence ECD.
We were able to evaluate some relationship between phosphate and ECD by including analysis of plasma ET-1 and ADMA. We found that phosphate was independently associated with ET-1, an endothelial cell derived vasoconstrictor. A phosphate-induced increase in ET-1 has been previously observed in both human endothelial cells and rat models of CKD[20]. Our novel finding in humans warrants longitudinal research to determine if improving phosphate control lowers ET-1 and possibly TPRI and BP. We also found that serum phosphate and PTH were associated with ADMA, and endogenous inhibitor of nitric oxide synthase that is associated with cardiovascular morbidity and mortality in HD patients[5]. This is consistent with the findings of Coen et al. revealing phosphate and PTH are individually associated with ADMA in ESRD patients[21].
Overall, our findings demonstrate a significant relationship between serum phosphate, vasoconstriction, and ECD in HD patients. It will ultimately need to be determined whether improvement in serum phosphate alone would be sufficient to improve ECD and lower BP in this population. One small randomized trial in CKD IV patients found that treatment with the phosphate binder sevelamer improved FMD, but use of a calcium-containing binder calcium acetate did not[22]. Improvement in FMD was strongly associated with levels of a calcification inhibitor, fetuin A, bringing forth an additional player in the interaction between MBD and vascular health. In that study, the phosphate remained high after treatment and many patients with comorbidities found in CKD and HD patients (diabetes, coronary artery disease, smoking, and use of renin-angiotensin-aldosterone system inhibitors or statins) were excluded. Another trial in a more heterogeneous CKD population found no significant changes or between-group differences in pulse wave velocity, coronary artery calcium score, or reactive hyperemia index (to assess ECD) in the participants randomized to lanthanum, calcium acetate, or low phosphate diet[23]. These studies highlight the need for further research that comprehensively takes into account the numerous mediators of MBD on vascular function and BP.
Limitations to our study include its relatively small size and the inability to draw conclusions about causality due its observational nature. Furthermore, we used ET-1 and ADMA as biomarkers for ECD. Because this was a retrospective analysis, we did not have alternative measurements such as FMD available. Additionally, the serum phosphate and other HD lab measurements were obtained in the context of routine clinical care and did not occur on the exact same date as our BP and TPRI measurements. However, we used the most recent measurements of phosphate preceding our measurements that usually occurred within a 1–2 week period. Our study had numerous strengths related to the variables that we ascertained and controlled for in the analysis to establish the presence of an independent association.