ORIGINAL ARTICLE

Dietary intake, blood pressure andosteoporosis

J Woo1, T Kwok1, J Leung2 and N Tang3

1Department of Medicine and Therapeutics, The Chinese University of Hong Kong, Shatin, NT, Hong Kong;2Jockey Club Centre for Osteoporosis Care and Control, The Chinese University of Hong Kong, Shatin,NT, Hong Kong and 3Department of Chemical Pathology, The Chinese University of Hong Kong, Shatin,NT, Hong Kong

Both hypertension and osteoporosis have commonunderlying nutritional aetiology, with regards to dietarycations intake. We tested the hypothesis that sodiumintake reflected in urinary Na/Cr and blood pressurewould be negatively associated with bone mineraldensity (BMD), whereas other cations may have oppo-site associations. Subjects were part of a study of bonehealth in 4000 men and women aged 65 years and over.A total of 1098 subjects who were not on antihyperten-sive drugs or calcium supplements and who providedurine samples were available for analysis. Logisticregression was used to examine associations betweentotal hip and lumbar spine BMD, age, gender, bodymass index (BMI), urinary Na/Cr, K/Cr, calcium andmagnesium intake, systolic blood pressure and diastolic

blood pressure. Total hip BMD was inverselyassociated with age, being female and urinary Na/Cr,and positively associated with BMI, urine K/Cr anddietary calcium intake. Lumbar spine BMD was inverselyassociated with being female and urinary Na/Cr,and positively associated with BMI, dietary calciumintake and SBP. We conclude that sodium intake,reflected by urinary Na/Cr, is the major factor linkingblood pressure and osteoporosis as shown by theinverse relationship with BMD. The findings lend furtheremphasis to the health benefits of salt reduction in ourpopulation both in terms of hypertension and osteo-porosis.Journal of Human Hypertension (2009) 23, 451–455;doi:10.1038/jhh.2008.156; published online 18 December 2008

Keywords: blood pressure; osteoporosis; sodium; calcium; magnesium; potassium

Introduction

The contribution of nutritional factors to the devel-opment of hypertension is well recognized, inparticular with reference to the four cations sodium,potassium, calcium and magnesium,1–6 both interms of nutrient intake quantities as well asfrequency of consumption of food group.7,8 The linkbetween hypertension and osteoporosis have notreceived as much attention, in terms of commonunderlying pathogenetic mechanism, perhaps asboth conditions are common with ageing. Onepathway would be the increased obligatory urinarycalcium excretion with increased urinary sodiumexcretion, which occurs with increased dietarysodium intake. This has been confirmed in theHong Kong Chinese population.9 Increased urinarycalcium loss predisposes to increased bone loss andosteoporosis. A direct connection may also occur

through the calcium-dependent vasoactive proper-ties of calcium-regulating hormones, which arelinked to the activity of the rennin–angiotensinsystem.10 Salt and calcium intake exert reciprocaleffects on these two hormone systems and on bloodpressure. For example, a recent population studyshowed an inverse relationship between systolicblood pressure (SBP) and serum 25 hydroxyvitaminD concentrations.11 Both SBP and diastolic bloodpressure (DBP) were higher in patients with lowcalcium intake and high serum parathyroidhormone concentration12–14 and age-related rise inSBP was inversely associated with calcium intake.15

However, a randomized controlled trial of calciumsupplementation in older New Zealand womenshowed only a small and transient hypotensiveeffect in most women, the effect being more markedin those with a calcium intake of o600 mg day�1,16

and no effect in normotensive or hypertensivepeople in another study.17 A recent CochraneDatabase analysis concluded that the causal rela-tionship between calcium supplementation onblood pressure is weak.18

Earlier studies in Chinese populations showed apositive association between sodium intake,19 but aninverse association with calcium intake was only

Received 2 September 2008; revised 5 November 2008; accepted 5November 2008; published online 18 December 2008

Correspondence: Professor J Woo, Department of Medicine andTherapeutics, The Chinese University of Hong Kong, 9/F,Clinical Sciences Building, Prince of Wales Hospital, Shatin, NT,Hong Kong.E-mail: jeanwoowong@cuhk.edu.hk

Journal of Human Hypertension (2009) 23, 451–455& 2009 Macmillan Publishers Limited All rights reserved 0950-9240/09 $32.00

www.nature.com/jhh

observed in older vegetarian women, among whomthe prevalence of osteoporosis is high.20 In a study ofbone health, among 4000 men and women aged 65years and over living in the community, we took theopportunity to examine the relationship betweenblood pressure, bone mineral density (BMD) andintake of calcium, sodium, potassium and magne-sium. We hypothesize that sodium intake as re-flected by urinary Na/Cr and blood pressure wouldbe negatively associated with BMD, while othercations may have opposite associations.

Materials and methods

A total of 2000 men and 2000 women aged 65 yearsand over living in the community were invited toattend a health check carried out in the School ofPublic Health of the Chinese University of HongKong, by placing recruitment notices in communitycentres for the elderly and housing estates. Severaltalks were also given at these centres explaining thepurpose, procedures and investigations to be carriedout. Subjects were volunteers, and the aim was torecruit a stratified sample so that approximately33% were in each of these age groups: 65–69, 70–74,75þ years. The study was approved by the ClinicalResearch Ethics Committee of the Chinese Univer-sity of Hong Kong, which requires informed consentto be obtained.

A questionnaire containing information regardingdrug history and dietary intake was administered byan interviewer. Dietary intake was assessed using a7-day food frequency questionnaire, and man nu-trient quantitation per day was calculated usingfood tables derived from McCance and Widdow-son21 and the Chinese Medical Sciences Institute.22

The food frequency questionnaire consisted of itemsin the following seven categories: bread/pasta/rice(16 items); vegetables (63 items); fruits (26 items);meat (39 items)/fish (31 items)/eggs (5 items);beverages (37 items); dinsum/snacks (39 items);soups (10 items); and oil/salt/sauces. Items chosenwere those most frequently consumed, based onprevious local surveys. Each subject was asked tocomplete the questionnaire—the food item, the sizeof each portion, the number of times of consumptioneach day and each week. Portion size was explainedto subjects using a catalogue of pictures of indivi-dual food portions. The amount of cooking oil wasestimated according to the method of preparingdifferent foods: 0.2 tablespoon for steaming fish orstir frying half a portion of vegetables, and 1tablespoon for stir frying one portion vegetables orone portion of meat.23

Blood pressure was measured after 5 min rest inthe sitting position using a standard mercurysphygmomanometer (WA Baum Co. Inc., Copiague,NY, USA). The first and fifth Korotkoff phases wererecorded as SBP and DBP. One reading was taken.Body weight was measured, with subjects wearing a

light gown, by the Physician Balance Beam Scale(Healthometer, Chicago, IL, USA). Height wasmeasured by the Holtain Harpenden stadiometer(Holtain Ltd, Crosswell, UK). Body mass index(BMI) (weight in kg divided by the square of heightin m) was calculated. BMD at the hip and lumbarspine region was measured using dual energy X-rayabsorptiometry (Hologic 4500 W, software version11.2).

A fasting urine sample was collected for analysisof sodium (Na), potassium (K) and creatinine (Cr),and the ratios Na/Cr and K/Cr were used asindicators of dietary sodium and potassium intake.All urinary analytes were analysed in the routinehospital laboratory on a Roche service analyzer.Urine Na and K were measured by the principle ofindirect Ion Selective Electrode and the analysiscoefficient of variation (CVs) were 2.2 and 1.6%respectively. Urine creatinine was measured bykinetic colourimetric assay based on modified JaffeReaction, and CV of the assay was about 2.0%.

Statistical analysisSubjects who were on hypertensive medication orcalcium supplements or who did not provide a urinesample were excluded from statistical analysis.Multiple logistic regression was carried out withtotal hip BMD as dependent variable. Independentvariables were age, female gender, BMI, urine Na/Crand K/Cr ratio, magnesium and calcium intake, andSBP and DBP. The analysis was repeated withlumbar spine BMD.

Results

Of the 4000 subjects, after excluding those usingantihypertensive medication and calcium supple-ments (n¼ 2061) and those without urine samples(n¼ 963), data from 1098 subjects were available foranalysis (809 women, 289 men). Total hip BMD wasinversely associated with age, being female andurinary Na/Cr, and positively associated with BMI,urine K/Cr and dietary calcium intake. Together thevariables account for 37% of the variance (Table 1).Urine Na/Cr, K/Cr and calcium intake have similarmagnitude of association with BMD, being less thanage, gender or BMI. No association was observedbetween SBP or DBP and BMD. Lumbar spine BMDwas inversely associated with being female andurinary Na/Cr, and positively associated with BMI,dietary calcium intake and SBP (Table 2). Thesevariables account for 27% of the variance. Dietarymagnesium intake was not associated with BMD ateither site.

Discussion

In our Chinese population, an earlier populationsurvey showed a high sodium and low calcium

Diet, blood pressure and osteoporosisJ Woo et al

452

Journal of Human Hypertension

intake.24 Only 22% of the population had an intakeof sodium p2300 mg day�1, while the averagecalcium intake was 570 mg day�1. The finding ofthis study would be indicative of associations for apopulation with high sodium and low calciumintake. Both low hip and spine BMD were associatedwith higher sodium intake, as indicated by urinaryNa/Cr ratio. However, this adverse effect may beameliorated by higher calcium intake and alsohigher intake for potassium (as indicated by higherurinary K/Cr) for hip BMD. This finding supportsthe pathway of increased obligatory calcium excre-tion with increased sodium excretion as a result ofincreased dietary sodium intake, as the commonunderlying pathogenetic mechanism for both hyper-tension and osteoporosis.

Furthermore, BMD appeared to be more correlatedwith Na/Cr and K/Cr compared with BP, as onewould expect the same associations to be presentbetween BP and BMD, but no significant associationwas observed between BP and hip BMD. Althoughan association between SBP and lumbar spine BMDwas observed, the direction of the association waspositive instead of negative. Although the explana-tion for this observation is unclear, salt intake

reflected by urinary Na/Cr would not be a factor asit is negatively associated with BMD. Therefore it ispossible that in this population, the adverse effect ofhigh salt intake on BMD could be greater, if not atleast as great as that on BP. In this setting, increasingcalcium intake would be all the more important.

This study also highlights the beneficial role ofpotassium and magnesium in relation to hyperten-sion. Increasing potassium intake would also bebeneficial for bone health. The major food source ofmagnesium and potassium are vegetables, fruits anddairy products, suggesting that the finding from theDASH study using a dietary approach to hyperten-sion with fruits and vegetables among the majorcomponents may also be relevant for bone health.

There are several limitations in this study. Thereare inherent inaccuracies in nutrient quantitationfrom food frequency questionnaires, and casualurine collection (rather than 24 h urine collection)as indicators of sodium and potassium intake.Estimation of sodium intake is particularly difficultin this population owing to the liberal use ofadded salt, soya sauces and monosodium glutamatein preparing food. The 24 h urine collection quanti-fying sodium excretion is the accepted method in

Table 1 Regression of total hip BMD (g cm�2) (n¼ 1098)

Variable Mean (s.d.)/prevalence Unit Percent difference per unit (95% CI)

Age 72.3 (5.3) 5.3 �3.5 (�4.4, �2.6)Female 73.7% No/yes �16.1 (�18.2, �14)BMI (kg m�2) 23.2 (3.6) 3.6 8.6 (7.8, 9.5)Urine Na/Cr (mmolmmol�1) 17.3 (10.4) 10.4 �1.2 (�2.1, �0.3)Urine K/Cr (mmolmmol�1) 7.8 (3.3) 3.3 1.2 (0.2, 2.1)Dietary magnesium per day (mg) 379.9 (195.9) 195.9 0.1 (�0.9, 1)Dietary calcium per day (mg) 579.6 (282.2) 282.2 1.7 (0.7, 2.7)SBP 141.7 (18.9) 18.9 �0.2 (�1.2, 0.9)DBP 77.7 (8.9) 8.9 0.1 (�1, 1.1)

R2¼0.3741

Abbreviations: BMD, bone mineral density; BMI, body mass index; CI, confidence interval; DBP, diastolic blood pressure; SBP, systolic bloodpressure.Bold values (Po0.05).

Table 2 Regression of lumbar spine BMD (g cm�2) (n¼ 1098)

Variable Mean (s.d.)/prevalence Unit Percent difference per unit (95% CI)

Age 72.3 (5.3) 5.3 0.8 (�0.4, 1.9)Female 73.7% No/yes �16.9 (�19.6, �14.2)BMI (kg m�2) 23.2 (3.6) 3.6 9.1 (8, 10.3)Urine Na/Cr (mmolmmol�1) 17.3 (10.4) 10.4 �1.4 (�2.6, -0.2)Urine K/Cr (mmolmmol�1) 7.8 (3.3) 3.3 0.8 (�0.5, 2.1)Dietary magnesium per day (mg) 379.9 (195.9) 195.9 0.5 (�0.8, 1.7)Dietary calcium per day (mg) 579.6 (282.2) 282.2 2.9 (1.6, 4.1)SBP 141.7 (18.9) 18.9 1.5 (0.1, 2.9)DBP 77.7 (8.9) 8.9 �1.2 (�2.6, 0.2)

R2¼0.2728

Abbreviations: BMD, bone mineral density; BMI, body mass index; CI, confidence interval; DBP, diastolic blood pressure; SBP, systolic bloodpressure.Bold values (Po0.05).

Diet, blood pressure and osteoporosisJ Woo et al

453

Journal of Human Hypertension

many epidemiological studies examining the rela-tionship between salt intake and blood pressure.25,26

However, collection of 24 h urine in elderly peopleis more prone to inaccuracies as the procedure isdependent on adequate manual dexterity andcognitive function. The volume collected tends tobe less than the true 24 h urine output. Therefore, acasual urine sample would be more feasible in alarge-scale epidemiological study of the elderly.Furthermore, studies using casual urine specimenshave yielded similar results to studies using 24 hurine collections in examining the relationshipbetween dietary cation intake and BP.27–29 We didnot carry out urine Ca/Cr measurements to showthat Na/Cr is associated with Ca/Cr. Blood pressurewas only measured once. However, we were able tostudy a large number of subjects with DEXAmeasurements, dietary intake and urinary measure-ments. In spite of the limitations, we conclude thatsodium intake, reflected by urinary Na/Cr, is themajor factor linking blood pressure and osteoporosisas shown by the inverse relationship with BMD. Thefindings lend further emphasis to the health benefitsof salt reduction in our population both in terms ofhypertension and osteoporosis.

Acknowledgements

This study received support from the Jockey ClubCharities Trust, the Research Grants Council of HongKong CUHK 4101/02M.

References

1 Suter PM, Sierro C, Vetter W. Nutritional Factors in thecontrol of blood pressure and hypertension. Nutr ClinCare 2002; 5: 9–19.

2 Hajjar I, Kotchen T. Regional variations of bloodpressure in the United States are associated withregional variations in dietary intakes: the NHANES-III data. J Nutr 2003; 133: 211–214.

3 Van Leer EM, Seidell JC, Kromhout D. Dietary calcium,potassium, magnesium and blood pressure in theNetherlands. Int J Epidemiol 1995; 24: 1117–1123.

4 Sacks FM, Brown LE, Appel L, Borhani NO, Evans D,Whelton P. Combinations of potassium, calcium, andmagnesium supplements in hypertension. Hyper-tension 1995; 26: 950–956.

5 Yang CY, Chiu HF. Calcium and magnesium indrinking water and the risk of death from hyper-tension. Am J Hypertens 1999; 12: 894–899.

6 INTERSALT. Intersalt: an international study of elec-trolyte excretion and blood pressure. Results for 24 hurinary sodium and potassium excretion. IntersaltCooperative Research Group. BMJ 1988; 297: 319–328.

7 Sacks FM, Svetkey LP, Vollmer WM, Appel LJ, BrayGA, Harsha D et al. Effects on blood pressure ofreduced dietary sodium and to Dietary Approachesto Stop Hypertension (DASH) diet. DASH-sodiumCollaborative Research Groups. N Engl J Med 2001;344: 3–10.

8 Akita S, Sacks FM, Svetkey LP, Conlin PR, Kimura G.Effects of the dietary approaches to stop hypertension(DASH) diet on the pressure-natriuresis relationship.Hypertension 2003; 42: 8–13.

9 Ho SC, Chen YM, Woo JL, Leung SS, Lam TH, JanusED. Sodium is the leading dietary factor associatedwith urinary calcium excretion in Hong Kong Chineseadults. Osteoporosis Int 2001; 12: 723–731.

10 Resnick LM. The role of calcium in hypertension.A hierarchal overview. Am J Hypertens 1999; 12: 99–112.

11 Judd SE, Nanes MS, Ziegler TR, Wilson PW, Tangpri-cha V. Optimal vitamin D status attenuates the age-associated increase in systolic blood pressure in whiteAmericans: results from the third National Health andNutrition Examination Survey. Am J Clin Nutr 2008;87: 136–141.

12 Jorde R, Sundsfjord J, Haug E, Bonaa KH. Relationbetween low calcium intake, parathyroid hormone, andblood pressure. Hypertension 2000; 35: 1154–1159.

13 Jorde R, Bonaa KH. Calcium from dairy products,vitamin D intake, and blood pressure: the TromsoStudy. Am J Clin Nutr 2000; 71: 1530–1535.

14 Schroder H, Schmelz E, Marrugat T. Relationshipbetween diet and blood pressure in a representativeMediterranean population. Eur J Nutr 2002; 41: 161–167.

15 Majjar IM, Grim CE, Kotchen TA. Dietary calciumlowers the age-related rise in blood pressure in theUnited States: the NHANES III survey. J Clin Hyper-tension 2003; 5: 122–126.

16 Reid IR, Horne A, Mason B, Ames R, Bava U, GambleGD. Effect of calcium supplementation on body weightand blood pressure in normal older women: arandomized controlled trial. J Clin Endocrinol Metab2005; 90: 3824–3829.

17 Nowson C, Morgan T. Effect of calcium carbonate onblood pressure in normotensive and hypertensivepeople. Hypertension 1989; 13: 630–639.

18 Dickinson HO, Nicolson DJ, Cook JV, Campbell F,Beyer FR, Ford GA et al. Calcium supplementation forthe management of primary hypertension in adults.The Cochrane Database of Systematic Reviews 2008; 4.AN:00075320-100000000-03558.

19 Woo J, Lau E, Chan A, Cockram C, Swaminathan R.Blood pressure and urinary cations in a Chinesepopulation. J Hum Hypertens 1992; 6: 299–304.

20 Kwok TC, Chan TY, Woo J. Relationship of urinarysodium/potassium excretion and calcium intake toblood pressure and prevalence of hypertension amongolder Chinese vegetarians. Eur J Clin Nutr 2003; 57:299–304.

What is known about topicK The prevalence of hypertension and osteoporosis increases

with age.K Sodium intake is a well-established risk factor for both

conditions.K Relative contributions of other cations such as calcium,

magnesium and potassium to both conditions are not asclear.

What this study addsK Sodium is the major factor linking blood pressure and

osteoporosis.K Calcium and potassium may ameliorate the adverse effect of

sodium on bone health.

Diet, blood pressure and osteoporosisJ Woo et al

454

Journal of Human Hypertension

21 Paul AA, Southgate DAT. McCance & Widdowson’s:The Composition of Foods, 4th edn, HM SO: London,UK, 1978.

22 Yang Y, Wang G, Pan G (eds). Institute of Nutrition &Food Safety, China CDC, In: China Food Composition2002. University Medical Press: Peking, China,2002.

23 Woo J, Leung SSF, Ho SC, Lam TH, Janus ED. A foodfrequency questionnaire for use in the Chinesepopulation in Hong Kong: description and examina-tion of validity. Nutr Res 1997; 17: 1633–1641.

24 Woo J. Nutrition and health issues in the general HongKong population. HK Med J 1998; 4: 383–388.

25 INTERSALT Cooperative Research Group. INTER-SALT: an international study of electrolyte excretion

and blood pressure. Results for 24 h urinary sodiumand potassium excretion. Br Med J 1988; 297: 319–328.

26 Kesteloot H, Huang DX, Li YL, Geboers J, Joossens JV.The relationship between cations and blood pressurein the People’s Republic of China. Hypertension 1987;9: 654–659.

27 Bruce NG, Cook DG, Shaper AG, Radcliffe JG. Casualurine concentrations of sodium, potassium and crea-tinine in populations studies of blood pressure. J HumHypertens 1990; 4: 597–602.

28 Khaw KT. Blood pressure and casual urine electrolytesin 93 London workers. Clin Sci 1983; 65: 243–245.

29 Khaw KT, Rose GA. A population study of bloodpressure and associated factors in St Lucia, WestIndies. Int J Epidemiol 1982; 11: 372–377.

Diet, blood pressure and osteoporosisJ Woo et al

455

Journal of Human Hypertension