European Journal of Clinical Nutrition (1998) 52, 29±33 ß 1998 Stockton Press. All rights reserved 0954±3007/98 $12.00

Phosphate supplementation in young men: lack of effect on calcium and turnover

A Whybro, H Jagger, M Barker and R Eastell

From the Department of Human and Clinical , and the Centre for , University of Shef®eld, Shef®eld, UK

Objective: To examine the effect of phosphate supplements on calcium homeostasis and bone turnover in young men. Design: Study 1 was a randomised, controlled, cross-over trial of 1000 mg elemental phosphate given for one week, with a standard diet of 800 mg=d each of calcium and . Study 2 was an escalating dose study of 0, 1000, 1500 and 2000 mg=d elemental phosphate, each given for one week, with a standard diet of 1000 mg=d each of calcium and phosphate. Setting: Northern General Hospital, Shef®eld. Subjects: Study 1, 10 healthy men ages 19±32 y. Study 2, 12 healthy men ages 19±38 y. Interventions: Sodium phosphate tablets each containing 500 mg elemental phosphorus (16.1 mmol=d), given with meals. Results: Study 1, administration of 32.2 mmol=d phosphorus resulted in a mean increase of 13.2 mmol=d urinary phosphorus, a mean decrease of 1.1 mmol=d urinary calcium, and a mean increase of 7 ng=mL in serum intact PTH. There were no changes in serum phosphate, osteocalcin or urinary N-telopeptide excretion. Study 2, administration of 64.4 mmol=d phosphorus resulted in a mean increase of 27.2 mmol=d urinary phosphorus, a mean decrease of 2.4 mmol=d urinary calcium, with no change in serum phosphate, PTH or urinary deoxypyridinoline. Conclusions: Phosphate supplementation of the diet does not affect bone turnover in young men. Sponsorship: Arthritis and Rheumatism Council Programme Grant. Descriptors: ; osteocalcin; deoxypyridinoline; phosphate

Introduction ever, estimated daily intake of calcium and phosphorus in adult men in Britain is 900 and 1450 mg=d, resulting in a Although phosphate is an essential necessary for molar ratio of 1.25 (Gregory et al, 1990). normal functioning, it has been suggested that relative to There is some support for an association between high the amount of calcium ingested, phosphate is consumed in phosphate to calcium ratios and an increased risk of amounts that are too high for optimal bone health (Calvo & osteoporosis. Forearm bone content has been Park, 1996; Calvo, 1993). Phosphate is present in most associated with an increase in this ratio (Tylavsky & , but the diet also includes phosphate compounds such Anderson, 1988; Lukert et al, 1987; Yano et al, 1985). as which are added during the manufactur- Radius BMC was related to phosphorus intake in preme- ing of foods such as frozen poultry, bacon, processed nopausal women (Metz et al, 1993), but this may have been cheeses, instant soups, desserts, sauces and carbonated due to a confounding effect of protein intake (Barker et al, beverages. Phosphates are used as pH regulators, emul- 1994). si®ers and anti-caking agents in order to reduce such Administration of a diet high in phosphate and low in problems as nutrient loss, fat oxidation and moisture calcium has resulted in an increase in serum parathyroid migration. As a result of the growing use and consumption hormone (Calvo et al, 1988) as has the administration of of phosphate-containing foods, individual phosphate in- phosphate salts to young subjects (Silverberg et al, 1986). takes have increased over the past ten years, although it The important question is whether these increases in para- is not yet known whether these levels will have any thyroid hormone levels result in increased signi®cant effects upon bone health (Gilbert et al, 1985). and hence accelerated bone loss. The recommended ratio of calcium to phosphorus in the There have been major advances in our ability to study diet is 1:1 (on a molar basis) resulting in recommended bone resorption with the introduction of assays based on the values of 700 mg=d of calcium and 550 mg=d phosphorus pyridinium crosslinks of type I collagen (Eastell, 1994) for adults in the UK (Department of Health, 1991). How- such as deoxypyridinoline (speci®c to bone) or N-telopep- tide crosslink of type I collagen. The present study was set up to use these new tools. Correspondence: Prof R Eastell, Department of Human Metabolism and The study was designed in two parts. Firstly, to evaluate Clinical Biochemistry, Clinical Sciences Centre, Northern General the effects of low and high phosphorus, subjects were given Hospital, Herries Road, Shef®eld S5 7AU, UK. Received 7 July 1997; revised 27 August 1997; accepted 10 September in random order these diets to evaluate their effect on bone 1997 turnover. Secondly, to evaluate the effects of extremes of Phosphate supplements and bone turnover A Whybro et al 30 phosphate intake, subjects were given escalating doses of Nichols Institute, San Juan Capistrano, CA). Urinary pyr- phosphate. This approach was taken because phosphate idinium crosslinks and were measured by HPLC after acid supplements commonly cause diarrhoea, especially when to allow measurement of total deoxypyridinoline introduced at high dose. (Colwell et al, 1993). The assays for PTH and pyridinium crosslinks were done in duplicate and the mean calculated. No subjects in either protocol suffered from disease known Methods to affect . None were smokers and all Study 1 Randomised crossover study were asked to abstain from alcohol during the protocols. A total of 10 healthy men ages 19±32 followed their usual Both protocols were approved by the Northern General diet for a week. Their habitual intake was assessed from a Hospital Ethics Committee and all subjects gave written 7 d weighed diet diary. Each subject was provided with informed consent. SOEHNLE electronic weighing scales with a range and accuracy of 0±2500 g and 2 g, respectively. They were Statistics provided with a diary for dietary recording and given In study 1, the biochemical data on the two diets were verbal tuition and written instructions on how to complete compared using the paired t-test. In study 2, the comparison an accurate food diary and use the weighing scales. They was made using two-way analysis of variance (factors were were provided with information about recording foods subject and diet) using the Scheffe test for comparison when eating out of the house and for recording homemade between phosphate doses. A P-value of 0.05 was taken as recipes and retail foods. Food preferences were also signi®cant. Statistical analysis was performed using Stat- obtained to enable the design of the test diet. The diaries graphics for Windows version 3 (Manugistics, Rockville, were analysed using COMP-EAT (Nutritional Services Ltd, MD, USA). London, England). Subjects were maintained on a dietary regimen for a two week period, which provided daily intakes of calcium of Results 800 mg and phosphate 800 mg. These intakes were achieved by providing subjects with evening meals and Study 1 Randomised crossover study asking them to consume prescribed foods for breakfast, Nine subjects completed the protocol. From the 7 d lunch and snacks. They were randomised to receive a weighed intake, the mean dietary calcium intake was phosphate supplement in the ®rst or the second of these estimated at 987 (223) mg=day (mean (s.d.)) and the two weeks. The supplement was 1000 mg=d phosphorus major sources were milk, bread, cheese and yoghurt. The given as two tablets of sodium acid phosphate 1.936 g, mean dietary phosphorus intake was 1538 (434) mg=d and sodium bicarbonate 350 mg, and potassium bicarbonate the major sources were milk, yoghurt, wholemeal bread, 315 mg, equivalent to 16.1 mmol phosphate, 20.4 mmol cheese, and tuna, and processed foods such as canned sodium and 3.1 mmol potassium (Phosphate Sandoz, soups, dried pasta mix and baked beans. Sandoz Pharmaceuticals, Camberley, Surrey, England). The urinary phosphate was higher on the high phosphate The two tablets were taken in divided doses with meals. diet (49.5, 11.6 mmol=d) than on the low phosphate diet was taken 3 h after breakfast and two 24 h urine (34.3, 8.3, P < 0.001) (Figure 1). The urinary calcium was collections were made at the end of each week. The lower on the high phosphate diet (4.4, 1.2 mmol=d) than on following measurements were made: serum calcium, the low phosphate diet (5.5, 1.3, P < 0.05). The intact PTH serum phosphate, urine calcium, phosphate and creatinine was higher on the high phosphate diet (33.7, 11.7 ng=mL) (by standard autoanalyser methods) and serum intact PTH than on the low phosphate diet (26.7, 10.9, P < 0.01). The by immunoradiometric assay (Allegro intact PTH) and serum phosphate did not differ on the high phosphate diet osteocalcin by radioimmunoassay (both from Nichols Insti- (1.38, 0.17 mmol=L) as compared to the low phosphate diet tute, San Juan Capistrano, CA) and urine N-telopeptides of (1.35, 0.16). The serum calcium did not differ on the high type I collagen (NTx) by ELISA (Ostex, Seattle, WA) and phosphate diet (2.37, 0.06 mmol=L) as compared to the low the results given in bone collagen equivalents (BCE) per phosphate diet (2.40, 0.05). day. All immunoassays were done in duplicate and the The serum osteocalcin did not differ on the high mean calculated. phosphate diet (7.0, 1.4 ng=mL) as compared to the low phosphate diet (6.6, 1.5). The urinary excretion of NTx Study 2 Escalating dose study did not differ on the high phosphate diet (532, 257 A total of 12 healthy men ages 19±38 y were maintained on nmol BCE=d) as compared to the low phosphate diet their habitual phosphate intake for a week. The habitual (540, 267). intake was assessed from a 7 d weighed diet diary as above. They were then given food containing 1000 mg=d calcium Study 2 Escalating dose study and phosphorus for four weeks. No supplement was given A total of eleven of the twelve subjects completed the for the ®rst study week and then 1000, 1500, and protocol successfully. One volunteer completed the study 2000 mg=d elemental phosphorus was given on each suc- up to the 1500 mg=d dose, but developed diarrhoea on the cessive week, with the dose split between each of the three 2000 mg=d dose and so his results from this last period meals of the day. The supplement was the same as for study were not used. Body weight over the study did not change one. signi®cantly. The tablets were counted at the end of the Blood was taken before breakfast and two 24 h urine study and compliance was above 90% in all cases. collections were made at the end of each week. The There were signi®cant changes in urinary phosphate and following measurements were made: serum calcium, calcium (Figure 2). The urinary phosphate was higher on serum phosphate, urine calcium, phosphate and creatinine the highest phosphate diet (61.4, 13.2 mmol=d) than on the (by standard autoanalyser methods) and serum intact PTH lowest phosphate diet (34.2, 12.4) (Figure 2). The urinary by immunoradiometric assay (Allegro intact PTH from calcium was lower on the highest phosphate diet (2.5, Phosphate supplements and bone turnover A Whybro et al 31

Figure 1 Changes in urinary phosphate, calcium and serum intact PTH in response to low (800 mg=d) and high (1800 mg=d) phosphate intakes in Study 1. There was an increase in urinary phosphate (P < 0.001, paired t test), a decrease in urinary calcium (P < 0.05), and an increase in PTH (P < 0.01).

1.4 mmol=d) than on the lowest phosphate diet (4.9, 1.8). compared to the lowest phosphate diet (1.18, 0.17). The The intact PTH did not differ on the highest phosphate diet serum calcium did not differ on the highest phosphate diet (31.2, 14.0 ng=mL) as compared to the lowest phosphate (2.43, 0.07 mmol=L) as compared to the lowest phosphate diet (26.7, 7.6) (Figure 3). The serum phosphate did not diet (2.45, 0.06). differ on the highest phosphate diet (1.18, 0.14 mmol=L) as The urinary excretion of Dpd did not differ on the high phosphate diet (91, 43 nmol=d) as compared to the low phosphate diet (72, 41) (Figure 3).

Figure 2 Changes in urinary phosphate and calcium in response to escalating doses of phosphate (1000±3000 mg=d) in Study 2. There was an increase in urinary phosphate (P < 0.001, ANOVA) and a decrease in urinary calcium (P < 0.001). The thick line represents the mean and the error bars, the standard error of the mean. Time points that did not differ Figure 3 Changes in serum PTH and urinary deoxypyridinoline (Dpd) in (at P < 0.05) have the same letter (Scheffe test), namely urinary phosphate response to escalating doses of phosphate (1000±3000 mg=d) in Study 2. excretion was increased and urinary calcium decreased compared to There was no signi®cant change in either measurement. The thick line baseline after increasing the dose to 2500 and 3000 mg=d. represents the mean and the error bars, the standard error of the mean. Phosphate supplements and bone turnover A Whybro et al 32 Discussion in which phosphorus is present in the diet may have differing effects on calcium homeostasis (Zemel & Links- Both of these studies showed an increase in urinary excre- wiler, 1981). Orthophosphate administration may not cause tion of phosphate and a decrease in urinary calcium. The bone loss. In balance experiments, orthophosphate did not increase in urinary phosphate equivalent to about half of the have the adverse effects of (Zemel & administered dose is to be expected (Bell et al, 1977; Van Linkswiler, 1981). In the treatment of postmenopausal Den Berg et al, 1980) as the fractional absorption of osteoporosis, orthophosphate given along with etidronate phosphate is about 50%. The decrease in urinary calcium had no adverse (or bene®cial) effect (Watts et al, 1990). In is a result of decreased calcium absorption and an increase a small group of women with osteoporosis calcium balance in endogenous faecal calcium (Heaney & Recker, 1982). and radius BMD improved with orthophosphate therapy The decrease in calcium absorption results from a decrease (Goldsmith et al, 1976). If the present study had been in the active form of , (Portale et al. conducted with polyphosphates, the outcome could have 1986; Van Den Berg et al, 1980). Serum phosphate is an differed. important regulator of 1-alpha hydroxylase in the . Although the serum phosphorus at a single time point did not increase with phosphate supplements, it is likely that we would have found it to be increased had we included a Conclusions circadian rhythm study of serum phosphorus in our proto- col (as did Calvo et al, 1988). It is likely that changing from a low to a high dietary The serum intact PTH increased in the ®rst study by a phosphate diet has little effect on bone turnover in young mean of 26% and this is in keeping with the effect of high men and so is unlikely to affect bone mass at maturity. It phosphate (low calcium) diets in young men (Calvo et al, remains an open question as to whether it is an important 1988) and with the effect of phosphate supplementation determinant of peak bone mass in women or has an effect (Silverberg et al, 1986). The effect of phosphate appears to on bone loss in older women. be greater in women than in men (Calvo et al, 1988) and in older compared to younger women (Silverberg et al, 1989). References There was no effect of phosphate on PTH in the second Barker M, Cadogan J, Jones N & Eastell R (1994): Calcium, protein, and study. This was most likely due to the timing of the blood phosphorus intakes and bone density. Am. J. Clin. Nutr. 60, 455±456. sampling, other studies have shown higher levels of serum Bell RR, Draper HH, Tzeng DY, Shin HK & Schmidt GR (1977): phosphate and PTH after a meal (Calvo et al, 1990; Calvo Physiological responses of human adults to foods containing phosphate et al, 1988; Portale et al, 1989). The dose of phosphate used additives. J. Nutr. 107, 42±50. in the second study was high, and was above the 2.5th Calvo MS, Kumar R & Heath H (1988): Elevated secretion and action of serum parathyroid hormone in young adults consuming high phos- percentile reported for British men of 2310 mg=d (Gregory phorus, low calcium diets assembled from common foods. J. Clin. et al, 1990). Endocrinol. Metab. 66, 823±829. There was no change in bone resorption markers in Calvo MS, Kumar R & Heath H (1990): Persistently elevated parathyroid either study. The NTx assay is clearly increased in primary hormone secretion and action in young women after four weeks of ingesting high phosphorus, low calcium diets. J. Clin. Endocrinol. hyperparathyroidism and returns to normal with effective Metab. 70, 1334±1340. treatment (Guo et al, 1996). Total deoxypyridinoline is the Calvo MS (1993): Dietary phosphorus, calcium metabolism and bone. gold standard of bone resorption markers in that it is J. Nutr. 123, 1627±1633. speci®c to bone, correlates well with estimates of bone Calvo MS & Park YK (1996): Changing phosphorus content of the U.S. resorption by calcium kinetics (Eastell et al, 1997) and is diet: potential for adverse effects on bone. J. Nutr. 126, 1168S±1180S. Colwell A, Russell RG & Eastell R (1993): Factors affecting the assay of not affected by renal metabolism, unlike other crosslink urinary 3±hydroxy pyridinium crosslinks of collagen as markers of bone forms (Colwell & Eastell, 1996). Previous studies have resorption. Eur. J. Clin. Invest. 23, 341±349. reported no change in free deoxypyridinoline (Karkkainen Colwell A & Eastell R (1996): The renal clearance of free and conjugated & Lamberg-Allardt, 1996) and no change (Zemel & Links- pyridinium crosslinks of collagen. J. Bone Miner. Res. 11, 1976±1980. Cosman F, Shen V, Herrington B & Lindsay R (1991): Response of the wiler, 1981; Calvo et al, 1990; Silverberg et al, 1986) or an parathyroid gland to infusion of human parathyroid hormone- increase in urinary hydroxyproline (Calvo et al, 1988; Bell (1-34)hPTH-(1-34)idemonstration of suppression of endogenous et al, 1977). Hydroxyproline is a poor marker of bone secretion using immunoradiometric intact PTH-(-84) Assay. J. Clin. resorption and is affected by the collagen composition of Endocrinol. Metab. 73, 1345±1351. the diet (Colwell et al, 1993). Department of Health (1991): Dietary Reference Values for Food, Energy and for the United Kingdom, pp 150±151, London: HMSO. An increase in PTH would be expected to increase bone Eastell R, Colwell A, Hampton L, Reeve J (1997): Biochemical markers of resorption. For example, Cosman et al (1991) reported an bone resorption compared with estimates of bone resorption from increase in collagen crosslinks shortly after starting an radiotracer kinetic studies in osteoporosis. J Bone Miner Res 12, infusion of PTH. It may be that the increase in serum 59±65. Eastell R (1994): Biochemical markers. Spine: State Art Rev. 8, 155±170. phosphorus antagonises the effect of PTH on bone resorp- Gilbert L, Wenlock RW & Buss DH (1985): Phosphorus in the British tion. In organ culture experiments it has been shown that household food supply. Hum. Nutr. Appl. Nutr. 39, 208±212. the phosphate concentration modulates bone resorption. A Goldsmith RS, Jowsey J, Dube WJ, Riggs BL, Arnaud CD & Kelly PJ high media phosphate inhibits bone resorption. For exam- (1976): Effects of phosphorus supplementation on serum parathyroid ple, Yates et al (1991) reported that a high phosphate hormone and bone morphology in osteoporosis. J. Clin. Endocrinol. Metab. 43, 523±532. medium decreased production of new osteoclasts from Gregory J, Foster K, Tyler H & Wiseman M (1990): The Dietary and their precursors (mouse bone marrow culture) and Nutritional Study of British Adults pp 153±176. London: HMSO. decreased bone resorption by mature osteoclasts (chicken Guo CY, Thomas WE, Al-Dehaimi AW, Assiri AM & Eastell R (1996): and rat osteoclasts on dentine slices). Longitudinal changes in bone mineral density and bone turnover in postmenopausal women with primary hyperparathyroidism. J. Clin. Phosphorus is present in a wider range of foods than Endocrinol. Metab. 81, 3487±3491. calcium and it is added to food (usually as a polyphosphate) Heaney RP & Recker RR (1982): Effects of nitrogen, phosphorus, and and to cola drinks (as ). The different forms caffeine on calcium balance in women. J. Lab. Clin. Med. 99, 46±55. Phosphate supplements and bone turnover A Whybro et al 33 Karkkainen M & Lamberg-Allardt C (1996): An acute intake of phosphate Tylavsky FA & Anderson JJ (1988): Dietary factors in bone health of increases parathyroid hormone secretion and inhibits bone formation in elderly lactoovovegetarian and omnivorous women. Am. J. Clin. Nutr. young women. J. Bone Miner. Res. 11, 1905±1912. 48, 842±849. Lukert BP, Carey M, McCarty B, Tiemann S, Goodnight L, Helm M, Van Den Berg CJ, Kumar R, Wilson DM, Heath H III & Smith LH (1980): Hassanein R, Stevenson C, Stoskopf M & Doolan L (1987): In¯uence of Orthophosphate therapy decreases urinary calcium excretion and serum nutritional factors on calcium-regulating hormones and bone loss. 1,25-dihydroxyvitamin D concentrations in idiopathic hypercalciuria. Calcif. Tissue Int. 40, 119±125. J. Clin. Endocrinol. Metab. 51, 998±1001. Metz JA, Anderson JJ & Gallagher PN Jr (1993): Intakes of calcium, Watts NB, Harris ST, Genant HK, Wasnich RD, Miller PD, Jackson RD, phosphorus, and protein, and physical-activity level are related to radial Licata AA, Ross P, Woodson GC, Yanover MJ, Mysiw J, Kohse L, Rao bone mass in young adult women. Am. J. Clin. Nutr. 58, 537±542. MB, Steiger P, Richmond B & Chesnut CH, III (1990): Intermittent Portale AA, Halloran BP, Murphy MM, & Morris RC Jr (1986): Oral cyclical etidronate treatment of postmenopausal osteoporosis. New intake of phosphorus can determine the serum concentration of 1,25- Engl. J. Med. 323, 73±79. dihydroxyvitamin D by determining its production rate in humans. Yano K, Heilbrun LK, Wasnich RD, Hankin JH & Vogel JM, (1985): The J. Clin. Invest. 77, 7±12. relationship between diet and bone mineral content of multiple skeletal Portale AA, Halloran BP, & Morris RC Jr (1989): Physiologic regulation sites in elderly Japanese-American men and women living in Hawaii. of the serum concentration of 1,25-dihydroxyvitamin D by phosphorus Am. J. Clin. Nutr. 42, 877±888. in normal men. J. Clin. Invest. 83, 1494±1499. Yates AJ, Oreffo RO, Mayor K & Mundy GR (1991): Inhibition of bone Silverberg SJ, Shane E, Clemens TL, Dempster DW, Segre GV, Lindsay R resorption by inorganic phosphate is mediated by both reduced osteo- & Bilezikian JP, (1986): The effect of oral phosphate administration on clast formation and decreased activity of mature osteoclasts. J. Bone major indices of skeletal metabolism in normal subjects. J. Bone Miner. Miner. Res. 6, 473±478. Res. 1, 383±388. Zemel MB & Linkswiler HM (1981): Calcium metabolism in the young Silverberg SJ, Shane E, de la Cruz L, Segre GV, Clemens TL, & Bilezikian adult male as affected by level and form of phosphorus intake and level JP (1989): Abnormalities in parathyroid hormone secretion and 1,25- of calcium intake. J. Nutr. 111, 315±324. dihydroxyvitamin D3 formation in women with osteoporosis. N. Engl. J. Med. 320, 277±281.