Consequences of a high phosphorus intake on mineral metabolism and bone remodeling in dependence of calcium intake in healthy subjects – a randomized placebo-controlled human intervention study

The association between elevated plasma phosphate concentrations and cardiovascular morbidity and mortality is well known in patients with chronic kidney disease (CKD) [1]. Nevertheless, in the last years, this association has been shown also for subjects without CKD and phosphate concentrations in the normal range. The Cholesterol And Recurrent Events (CARE) study as well as the Framingham Offspring Study showed a relation between serum phosphate concentrations >1.13 mmol/l and a higher risk for death and cardiovascular events in individuals free of CKD [2, 3]. However, the CARE and the Framingham Offspring Study have not determined the influence of dietary intake on serum phosphate.

In association with the phosphate metabolism in humans, the fibroblast growth factor 23 (FGF23) got more and more attention. Jüppner et Bergwitz [4] deemed FGF23 as a biomarker for the dysregulation of the phosphate balance in patients with CKD; this biomarker is probably involved in the pathophysiology of cardiovascular events in these patients. Together with vitamin D and parathyroid hormone (PTH), FGF23 equilibrates high plasma phosphate concentrations (after dietary intake) by increasing renal phosphate excretion and decreasing absorption of phosphate in the gut [5]. In the early stage of CKD, first FGF23 and then PTH increases to normalize serum phosphate concentration. But this mechanistic FGF23 increase is not yet fully understood [6]. Furthermore, the physiology of FGF23 in healthy adults after phosphate load is not sufficient examined.

Ritz et al. [7], Uribarri and Calvo [8] assumed, that the increased use of phosphorus (P) in food processing and the consequentially increased intake of P, which exceeds the nutritional recommendation, is responsible for the association with cardiovascular morbidity and mortality, even with serum phosphate concentrations in the normal range. Indeed, the recommendation of the US Institute of Medicine for adults between 19 and 50 years are 700 mg P/d, but the intake of P is far above this recommendation in the USA (approximately 1600 mg/d for men and 1200 mg/d for women) [9]. Due to the latest report of the European Food Safety Authority in Europe the phosphorus intakes ranges between 1000–1767 mg/d and are far above the recommendations, too [10]. Therefore, investigations of the consequences after high P intake in healthy humans is an important area of research.

Studies regarding high P intakes have some limitations, e.g., comparison of restricted versus supplemented P diet [11], short-term supplementation [12] and small sample size [13]. The present double-blind, placebo-controlled human intervention study in parallel design tries to compensate these former limitations and to complement the previous findings to get more insight and information about the consequences of a high P intake in healthy adults taking different calcium (Ca) intakes into consideration. The objectives of the present study are to determine the influence of high P intakes on phosphorus, calcium, magnesium and iron metabolism as well as FGF23 concentrations and bone remodeling marker.

The baseline characteristics of participants of the three intervention groups are presented in Table 1. There were no significant differences in age, BMI, 25(OH)D and CKD-EPI between intervention groups.

The epidemiological studies from Tonelli et al. [3] and Dhingra et al. [2] only determined an association between serum phosphate concentrations and the risk for cardiovascular diseases, but not the influence of the dietary P intake. Chang et al. [17] showed that high P intakes (>1400 g/d), but not fasting serum phosphorus concentrations are associated with an increased mortality in a healthy US population. The authors concede that additional studies are necessary to determine whether this relation is causal.

The results of the present study show clearly that additional intake of 1000 mg P/d did not influence fasting plasma phosphate concentrations of healthy adults after eight weeks of intervention, independent of Ca intake. This is in accordance with other studies, which examined among other things fasting phosphate concentrations [11, 18‐20]. By looking at short-term interventions, it should be noticed that serum phosphate concentrations increased after high P intake during the day, but returned to fasting concentrations [20‐24]. The phosphate homeostasis in healthy subjects comprises obviously a regulation of acute P loads. Since renal P excretions increased after additional P intakes in the present study, it can be hypothesized that elimination of P via the kidney occurred. Interestingly, the fecal P concentrations increased as well, which indicates the formation of amorphous calcium phosphate in the intestine [16, 25, 26]. Noteworthy, the renal excretion of Ca decreased after P1000/Ca0 intervention, while the renal Ca excretions remained stable in the Ca-supplemented intervention groups. This indicates a negative influence of a high P and a normal Ca diet on Ca balance and is in accordance with Kemi et al. [27], who reported a disturbed Ca metabolism by a Ca:P ratio below 0.5 (present study P1000/Ca0 0.4). However, the other two intervention groups with Ca and P (P1000/Ca500 and P1000/Ca1000) showed no negative effect on Ca balance since plasma Ca concentrations increased and renal excretions remained unchanged. Our results and the existing literature give evidence for maintenance of a positive Ca balance, a high P intake has to be accompanied by a high Ca intake [27‐29].

FGF23 is a major factor in the regulation of phosphate homeostasis, especially the regulation of blood phosphate after dietary intake [5]. Interestingly, the plasma FGF23 is discussed as a risk factor for mortality in patients with CKD [30]. Gutiérrez et al. [30] concluded that increased FGF23 could be a new biomarker for death in early CKD, when phosphate concentration is still normal. Mortality-associated FGF23 concentrations are far away the FGF23 concentration of healthy subjects in the present study. This is not surprising since patients with CKD have 3-fold higher FGF23 concentrations compared to healthy subjects [13]. In our study were five subjects with FGF23 concentrations above 300 kRU/l in at least one study period. These subjects showed regular P and Ca parameters and did not suffer from CKD. The results of the remaining 56 subjects were similar in the three intervention groups. The FGF23 concentration increased on average by 14 kRU/l (17 %) after four weeks of supplementation compared to placebo (only significant for P1000/Ca1000) and decreased by 19 kRU/L (33 %) after eight weeks compared to four weeks of supplementation. Burnett et al. [31] showed an increase of approximately 37 % for FGF23 (extracted from Fig. 2, taking the reported baseline value into account: mean increase of 21 kRU/l) within five days of P supplementation. Similar, but not significant, results found Ferrari et al. [20]. It seems that FGF23 temporarily increases after P load, but decreases in adaption to the diet. To the best of our knowledge, there is no study which determined FGF23 after intervention longer than four weeks. We could show for the first time that an increase of FGF23 is not a permanent effect. Therefore, we assume that other mechanism keeps phosphate concentration in homeostasis. Since PTH and 1,25(OH)₂D are the best known phosphate regulators it is necessary to focus these two. In the present study, PTH did not change after the three interventions. 1,25(OH)₂D increased in all intervention groups after eight weeks of supplementation compared to placebo, but significantly only in P1000/Ca0. This is little surprising since literature reports that high P intake results in PTH and FGF23 increases, which cause phosphaturia [29]. PTH stimulates and FGF23 inhibits 1,25(OH)₂D synthesis, resulting in an increase and decrease of P absorption, respectively. Additionally, FGF23 in turn inhibits PTH [29]. The unchanged fasting PTH concentrations after high P intake are in accordance with Grimm et al. [19] and Ix et al. [13], but not with other human intervention studies, which showed an increase in fasting PTH concentrations [18, 32‐34]. The main reason for the differences between the above-mentioned studies and the present one is probably the different intervention time (present study: 8 weeks; other studies: maximum 28 days).

The determined bone formation (osteocalcin and BAP) and resorption markers (CTX) decreased significantly in the Ca-supplemented groups. The P supplementation alone did not change the markers significantly. According to Coates et al. [35] the day-to-day variation should be taken into account for interpretation of bone markers, thus a significant change is generally determined as 2.8 times the biological variation. Due to the fact that the measured bone markers generally vary between 8–12 %, the change should be greater than approximately 30 % in order to get clinical relevance [35]. However, the determined markers of bone metabolism in the present study changed only approximately 20 %. Therefore, we could not observe a clinically relevant effect on bone remodeling after a high P intake in healthy subjects.

After P supplementation but without Ca supplementation, renal Mg excretion decreased. This outcome indicates a decreased absorption of Mg in the gut due to high P intake and is in accordance with Brink et al. [36]. These authors showed in an rat experiment that Mg forms insoluble complexes with Ca and phosphate. Greger et al. [37] showed that significantly less Mg was excreted via urine of men following high P consumption (2443 mg P) compared to moderate P diets (843 mg P). But the authors concluded, that the apparent Mg retention and thus the utilization of Mg was not affected. Transferrin concentration in plasma slightly increased due to  eight weeks of P1000/Ca500 intervention compared to placebo. The other iron metabolism marker did not change due to the interventions. An interaction of a calcium intervention with iron was shown in mostly short-term studies [38‐40]. In a long-term study of Minihane and Fairweather-Tait [41] no effect of iron metabolism after calcium supplementation was shown. The influence of phosphate intake on iron metabolism was not yet investigated. Thus, it should be taken into consideration, that a high P intake may unbalance Mg and iron metabolism.

Throughout the eight weeks of P supplementation, 16 % of subjects reported intestinal disturbances (diarrhea, flatulence, stomachache and/or constipation). This is a consequence of an osmotic effect in the intestine due to the high P concentration [19].

A potential limitation of the study is the determination of fasting blood parameters instead of postprandial ones, because it is known that phosphate concentrations change differently throughout the day, depending on the phosphors intake [22]. However, the above-mentioned epidemiological studies [2, 3, 17] used fasting measurements and the present study could show that eight weeks of high phosphorus intake could not change fasting phosphate concentration. Therefore, it should be taken into consideration that the higher risk for mortality and cardiovascular disease does not comes solely from the diet. It is most likely an effect of interaction between genetics, other not recognized diseases (e.g. kidney) and the diet.

Another limitation factor could be the potential declining compliance of the subjects after ten weeks of study. Nevertheless, due to the increased renal excretion as well as fecal concentration of P after eight weeks of P intervention the authors conclude a persistent compliance.

In all, our study showed that a high P intake did not influence fasting phosphate plasma concentrations in healthy adults. A high P intake without adequate Ca supplementation seems to have a negative impact on Ca metabolism. Thus, it can be stated that a well-balanced Ca:P ratio is an important prerequisite for a normal metabolism of Ca. There is only a temporary increase of FGF23 after P supplementation. Within eight weeks of a high P intervention, plasma concentrations of FGF23 remained within the normal range.