Bone Mineral Density in Chinese Elderly Female Vegetarians, Vegans, Lacto-Vegetarians and Omnivores

Bone mineral density in Chinese elderly female vegetarians, vegans, lacto-vegetarians and omnivores

EMC Lau1, T Kwok2, J Woo2 and SC Ho1

1Department of Community and Family Medicine, The Chinese University of Hong Kong, Hong Kong; and 2Department of Medicine, The Chinese University of Hong Kong

Objectives: To compare the bone mineral density and dietary intake of elderly Chinese vegetarian women with omnivores, to compare the bone mineral density of Chinese `vegans' and `lactovegetarians', and to study the relationship between nutrient intake and BMD in vegetarians. Design: A cross-sectional survey. Setting and subjects: A community-based study. The vegetarian women (aged 70±89 y) (n 76) were non- institutionalized subjects. All of them were Buddhists. Their bone mineral density were compared to normal elderly volunteers (aged 70±89 y) (n 109) who were recruited to establish normal BMD ranges. Their dietary intake was compared to omnivorous subjects from a previous dietary survey (n 250). Methods: Dietary assessment was by the 24 h recall method, and bone mineral density was measured by dual-X- ray-densitometry. The analysis of co-variance was used to compare the BMD between vegetarians and omnivores, with adjustment for potential confounders. The BMD in `vegans' and `lactovegetarians' were compared by similar methods. The t-test was used to compare dietary intake between omnivores and vegetarians. The relationship between nutrient intake and BMD was studied by correlation and multiple regression. Results: The dietary calorie, protein and fat intake were much lower, but the sodium=creatinine ratio was much higher in vegetarians than omnivores. The BMD at the spine was similar between vegetarians and omnivores. However, the BMD at the hip was signi®cantly lower in vegetarians at some sites (P < 0.05). There was no signi®cant difference in BMD between `vegans' and `lactovegetarians'. BMD in vegetarians appeared to be positively correlated with energy, protein and calcium intake; and negatively associated with urinary sodium= creatinine levels. Conclusions: There is a relationship between diet and BMD. The BMD at the hip was lower in vegetarians than omnivores, but no difference was observed between `vegans' and `lactovegetarians'. There is a complex relationship between the intake of various nutrient and BMD in vegetarians. Sponsorship: None Descriptors: Bone mineral density; vegetarians; vegans; lacto-vegetarians, omnivores

Introduction between vegetarians and omnivores (Tylavsky & Ander- son, 1988; Lloyd et al, 1991; Tesar et al, 1992; Hunt et al, Osteoporosis is rapidly becoming a major public health 1989; Marsh et al, 1988). problem in Hong Kong (Lau et al, 1990). Previous studies have indicated that dietary calcium intake is low in our population (Pun et al, 1989), and such low intakes were Subjects and methods shown to be associated with increased risk of both hip (Lau et al, 1988) and vertebral fracture (Chan et al, 1996). The The research protocol was approved by the Human practice of vegetarianism has a long history in the Chinese Research Ethics Committee of the Chinese University of culture, with a substantial number of elderly women Hong Kong. Informed consent were obtained from all abstaining entirely from animal products. In this group, subjects prior to the study. A cross-sectional survey was vegetarianism is usually practised for religious reasons. A conducted on 76 Chinese vegetarian women aged 70±89 y. study into the bone mineral density of vegetarian Chinese These women had all been vegetarians for over 30 y, and women may enable further understanding into the relation- all were ®t and ambulant. Women who consumed milk for ship between dietary intake and bone health. once a month or more were labelled as `lactovegarians', The bone mass in vegetarians have been studied in and their BMD was compared with the `vegans' or women Caucasians. In the past, the bone mineral content (BMC) who never consumed milk, in the latter part of the analysis. of small members of Caucasian vegetarians has been The 24 h recall method was used for assessing dietary reported to be higher than that of omnivores (Ellis et al, intake. The dietary assessment was administered by a 1972; Marsh et al; 1980; Marsh et al, 1983). However, in single trained interviewer. All subjects were asked to the more recent studies, the BMD was found to be similar recall the actual food and drink consumed, in the last 24 h. Pictures of average portion sizes of food were shown to subjects, and the amount consumed were recorded as a multiple (or fraction) of such sizes. The dietary intake Correspondence: Dr EMC Lau Received 7 July 1997; revised 26 September 1997; accepted 20 September of various nutrients was then calculated using Food Com- 1997 position Tables for South East Asia (Food and Agriculture Bone mineral density in Chinese elderly females EMC Lau et al 61 Organization, 1972). The dietary intake in vegetarians were Table 1 Life-style factors, anthropometric measurements and bone min- compared to values in Chinese omnivorous women, as eral density (BMD) of 76 Chinese vegetarian women and 109 omnivores obtained by similar methods in a previous dietary survey Vegetarians (n 76) Omnivores (n 109) (n 250). Mean (s.d.) Mean (s.d.) BMD at the hip and spine was measured by dual X-ray densitometry (Hologic). The C.V. for measurements at both Age 79.1 (5.2)* 77.0 (3.8) Height 1.45 (0.06) 1.47 (0.06) sites were 1% in our laboratory. The weight and height of Weight 48.8 (9.67) 49.9 (9.12) all subjects were also recorded. The BMD and anthropo- Cigarette smokers (%) 11 16 metric measurements of the vegetarian women were com- Alcohol drinkers (%) 1 2 pared to normal values obtained in omnivorous controls Percentage who walked 96% 97% outdoors every day (%) (n 109). Detailed dietary assessment had not been per- Spine 0.70 (0.13) 0.72 (0.14) formed in this control group. Femoral neck 0.49 (0.078)*** 0.53 (0.082) A fasting 20 ml specimen of urine was collected in a Intertrochantic area 0.69 (0.13) 0.73 (0.14) plain glass bottle, for estimating the sodium and creatinine Ward 0.29 (0.091)** 0.34 (0.094) levels. Urinary sodium was estimated by ¯ame photometry *P < 0.05 by t-test. using an IL-943 ¯ame photometer (Instrumentation **P < 0.05 *** P < 0.001 by ANCOVA, adjusting for age, height, weight, Laboratory Inc. Lexington, MA, USA). Urinary creatinine cigarette smoking and alcohol consumption. were estimated using the `Parallel' Analytic System (Amer- ican Monitor Corporation, Indiana, USA). The coef®cient of variation was 1% for both urinary sodium and creatinine. Table 2 Dietary intake in Chinese vegetarians and omnivores (as compared The dietary intake between the vegetarians and omnivores to Caucasian vegetarians) were compared by the analysis of co-variance (ANCOVA), Chinese Caucasian with adjustments for age, body weight, height, cigarette Chinese vegetarians omnivores vegetariansa smoking and alcohol consumption, and physical activity (as (n 76) (n 250) (n 28) measured as whether subjects walked outdoors every day). None of the study subjects performed recreational activity Median Mean (s.d.) Mean (s.d.) Mean (s.d.) regularly. The dietary intake in Chinese `vegans' and Energy (Kcal) 1112 1142 (340)* 1557 (510) 1652 (347) `lactovegetarians' compared by the t-test; while their Protein (g) 31 35 (16)* 60 (21) 63 (24) BMD was compared by ANCOVA, with adjustments for (% Kcal) 12 12 (3) 16 (5) 15 (4) Carbohydrate (g) 211 211 (61)* 248 (88) 242 (6) age, body weight, height, cigarette smoking and alcohol (% Kcal) 76 75 (9) 64 (12) 59 (8) consumption, and physical activity (as measured by Fat (g) 15 19 (14)* 34 (28) 56 (16) whether subjects walked outdoors every day). (% Kcal) 13 15 (9) 19 (12) 31 (7) To study the relationship between dietary intake and Calcium=1000 Kcal 278 338 (189)* 221 (168) 496 (210) Urinary Na=Cr ratio 23.3 28.4 (18.8)* 17.5 (4.3) NA BMD, correlation coef®cients between dietary intake and Urinary K=Cr ratio 7.0 7.8 (4.4)* 4.3 (4.5) NA BMD were calculated. Multiple regression was then performed, with the dietary intake as predictor variables and aData was adapted from Tesar et al, 1992. the BMD as the predicted variables. The dietary intake **P 0.001 by t-test between Chinese vegetarians and omnivores. from our study was compared to the results from a study by Tesar et al (1992). In their study, dietary intake was women (Tesar et al, 1992). The average energy consump- assessed as an average of 24 h record and 6 d dietary tion was signi®cantly lower in vegetarian women than in records. Such methods may yield slightly higher intake omnivores. For the vegetarian women, the main source of than the 24 h record alone. energy intake was from carbohydrates, with a much smaller percentage from fat and protein. The dietary calcium intake, as measured by calcium intake per 1000 Kcal, Results was higher in vegetarian subjects than omnivores. The BMD, anthropometric measurements and life-style factor of high salt intake among vegetarians was re¯ected in the vegetarians and omnivores very high urinary sodium=creatinine ratio. While the diet- The BMD, anthropometric measurements and life-style ary intake of calories, protein, fat and calcium was much factors of vegetarians are shown in Table 1. On the average, lower among Chinese vegetarians than Caucasian vegetar- the vegetarians were 2 y older than omnivores (P < 0.05 by ians, these were very similar between Chinese omnivores the t-test). However, there were no signi®cant difference in and Caucasian vegetarians. body weight, height, percentage of cigarette smokers, alcohol drinkers and subjects who walked outdoors Dietary intake and bone mineral density of Chinese between the vegetarians and omnivores. None of the lactovegetarians and Chinese vegans study subjects performed recreational activity other than The dietary intake and bone mineral density of lactovegetar- walking outdoors. While there was no difference in BMD at ians and vegans are shown in Table 3. The total dietary intake the spine between vegetarians and omnivores, the BMD at of energy, protein and fat were signi®cantly higher in vegans the hip (at the Femoral Neck and Ward's Triangle) were than lactovegetarians (P < 0.05). The dietary calcium intake much lower in the vegetarians. Such differences were was signi®cantly higher in the Chinese lactovegetarians signi®cantly different (P < 0.05) by ANCOVA. (P < 0.05). Milk intake in the lactovegetarians was mainly in the form of high calcium and low fat powdered milk. Dietary intake of Chinese vegetarians and Chinese omnivores Relationship between dietary intake and BMD In Table 2, the dietary intake of the Chinese vegetarian The correlation coef®cients and results of multiple regres- women is presented with the values of 250 Chinese sion are shown in Table 5. Dietary energy, protein and omnivores women, and values for Caucasian vegetarian calcium intake was positively correlated with BMD at all Bone mineral density in Chinese elderly females EMC Lau et al 62 Table 3 Dietary intake in Chinese vegans and lactovegetarians the dietary intake, as indicated by the R2, which ranged from 3.5% at the spine to 15% at the femoral neck Chinese Chinese vegans lactovegetarians respectively. (n 36) (n 40)

Median Mean (s.d.) Median Mean (s.d.) Discussion Energy (Kcal) 1241 1247 (364)** 1005 1048 (291) Protein (g) 35 39 (20)* 29 31 (11) The bone mass of vegetarians have been compared to (% Kcal) 12 12 (4) 11 12 (3) omnivores in Caucasian populations. The results of earlier Carbohydrate (g) 229 225 (59) 195 199 (61) studies suggested that vegetarians may have higher BMC (% Kcal) 73 73 (9) 79 76 (9) than omnivores (Ellis et al, 1972; Marsh et al, 1980; Marsh Fat (g) 21 24 (17)* 12 16 (11) (% Kcal) 15 16 (8) 11 14 (9) et al, 1983). However, in the more recent studies, the BMC Calcium=1000 Kcal 276 288 (101)* 282 382 (235) and BMD of vegetarians and omnivores were not found to be different. Two cross-sectional studies were conducted in *P 0.05. the late 1980s to compare the BMC of vegetarian and **P 0.01 by t-test comparing vegans and lactovegetarians. omnivores. In these studies, 88 subjects living in North Carolina (Tylavsky & Anderson, 1988) and 144 subjects Table 4 Anthropometric measurements and bone mineral density of living in California (Hunt et al, 1989) were investigated. vegans and lactovegetarians The mean calcium and calorie intakes were similar between Vegans Lactovegetarians vegetarians and omnivores, but mean protein intakes were (n 36) (n 40) signi®cantly lower in the vegetarians. The BMC at the Mean (s.d.) Mean (s.d.) forearm of vegetarians were found to be similar to omni- Age (y) 78.2 (4.9) 79.9 (5.4) vores. Moreover, no signi®cant difference was found in the Weight (Kg) 48.2 (9.6) 49.2 (9.8) spinal BMD of vegetarian and omnivorous pre-menopausal Height (m) 1.46 (0.06) 1.5 (0.06) women (Lloyd et al, 1991; Tesar et al, 1992). We have Cigarette Smokers (%) 3 18 studied the BMD of elderly Chinese vegetarian women, and Alcohol drinkers (%) 0 3 have compared these with the values from omnivorous Bone mineral density (g=cm2) controls. The BMD at the hip was found to be consistently Spine (L1±L4) 0.72 (0.15) 0.68 (0.11) Femoral neck 0.50 (0.08) 0.48 (0.08) lower in the vegetarians. Such differences could not be Intertrochanteric area 0.69 (0.12) 0.69 (0.14) explained by age or other life-style factors, for the differ- Ward triangle 0.29 (0.09) 0.29 (0.10) ence in BMD remained after adjusting for these variables. Before we attempt to account for the difference in BMD between vegetarians and omnivores by difference in dietary sites. However, the urinary sodium=creatinine ratios were intake, it is necessary to examine the method used for negatively correlated with BMD at the Intertrochanteric assessing intake. In this study, the 24 h recall method was Area and the Ward's Triangle. used to ascertain intake. Dietary assessment by the 24 h None of the dietary intake measurements was found to recall method may not be as accurate as weighed dietary be signi®cantly associated with BMD by multiple regres- records, since elderly subjects may have relatively poor sion. It is interesting to note the negative regression short term memory, and interday variations would not be coef®cients for dietary protein; and the positive values taken into account. However, our choice of dietary survey for urinary sodium creatinine ratios. A relatively small method was restricted by the methods used previously in proportion of the variation in BMD was accounted for by our control subjects. It has been demonstrated that 24 h

Table 5 Relationship between dietary intake and bone mineral density (in g=cm2) in Chinese vegetarians

(a) Correlation coef®cients Bone mineral density (g=cm2)

Dietary Intake=d Spine (L1±L4) Femoral neck Intertrochantric Area Ward's triangle Energy (Kcal) 0.09 0.20 0.19 0.11 Protein (g) 0.09 0.13 0.084 0.042 Calcium (g) 0.048 0.16 0.13 0.13 Urinary Na=Cr ratio 0.15 0.07 70.032 70.015

(b) Multiple regression models Predicted varible: bone mineral density (g=cm2) Predictor variables: dietary intake=d Spine (L1±L4) Femoral neck Intertrochanteric Area Ward's triangle

Age 78.1 6 1074 73.6 6 1073 76.6 6 1073* 76.1 6 1073** Energy (Kcal) 5.0 6 1075 6.7 6 1075 1.5 6 1074 6.4 6 1075 Protein (g) 7.9 6 1074 76.8 6 1074 73.6 6 1073 1:0 Â 10 3 Calcium (g) 5.4 6 1075 3.4 6 1075 1.5 6 1074 1.1 6 1075 Urinary Na=Cr ratio 1.2 6 1073 6.8 6 1074 4.9 6 1074 2.5 6 1074 Constant 0.67* 0.69** 1.08*** 0.72** R2 for model 0.035 0.12 0.15 0.15

Each column represented one multiple regression model. *P 0.05. **P 0.01. ***P 0.001. Bone mineral density in Chinese elderly females EMC Lau et al 63 recall was adequate in categorising subjects into levels of dietary calcium intake was high in subjects who consumed intake, and the use of this methods should enable valid milk, the total calories, protein and fat intake were lower in comparison in our study. the `lactovegetarians'. As we have contended, an adequate The difference in BMD between Chinese vegetarian and intake of calories and protein may be needed to optimize controls could be examined in the light of the difference in bone mass. Hence any bene®cial effect in the `lactovege- dietary intakes. Vegetarians had a lower energy intake, a tarians' due to a high calcium intake may be offset by a higher carbohydrate intake, a lower fat and a lower protein lower intake of protein and calories. intake than omnivores. Their dietary calcium and salt The complex relationship between diet and BMD was intake were also higher than in omnivores. While the well illustrated by our regression analysis. Bone mineral dietary intake of energy and protein of Chinese omnivores density seemed to be positively correlated with dietary were above the Recommended Dietary Allowance intake of energy, protein and calcium; and negatively (National Research Council, 1980) for East Asia, this was correlated with urinary sodium=creatinine ratio. However, not so in the vegetarians. In this group, both calories intake in multiple regression, protein appeared to be negatively (mean 1142 Kcal) and protein intake (mean 35 g) were associated with BMD. Due to the relatively small sample lower than the Recommended Dietary Allowance (National size, statistical signi®cance was not demonstrated. Never- Research Council, 1980) (which were 1400 Kcal and 37 g theless, it was clear that dietary intake accounted for only a respectively) in the elderly. fraction of the variation in BMD at various sites. The relationship between dietary protein intake and To quantify the relationship between dietary intake and bone mass remains controversial. Although some cross- BMD accurately, larger follow up studies are required. sectional studies suggest that a high protein intake may be Moreover, supplementation studies will provide the means detrimental. (Abelow et al, 1992; Hegsted & Linkswiler, for dietary recommendations in this unique population. 1981; Schuette & Linkswiler, 1982; Kerstetter & Allen, 1990; Geinez et al, 1993; Chevalley et al, 1991), an inade- References quate protein intake has also been shown to be detrimental to Abelow BJ, Holford TR & Insogna KL (1992): Cross-cultural association bone health. Protein supplementation has been shown to between dietary animal protein and hip fracture: A hypothesis. Calcif. prevent bone loss in osteoporotics (Tkatch et al, 1992). As Tissue Int. 50, 14±18. the protein intake in Chinese vegetarian were extremely low, Chan HHL, Lau EMC, Woo J, Lin F, Sham A & Leung PC (1996): Dietary bone loss may be excessive, hence resulting in lower BMD. calcium intake, physical activity and risk of vertebral fracture in Chinese. Osteoporosis Int. 6, 228±232. The results of meta-analysis (Cumming, 1990) and recent Chevalley T, Rizzoli R & Nydegger R, Slosman D, TKatch L, Rapin CH, clinical trials (Dawson Hughes et al, 1990; Prince et al, 1991) Vasey H & Bonjour JP (1991): Preferential low bone mineral density of demonstrated that a high dietary calcium in take has a the femoral neck in patients with a recent fracture of the proximal positive effect on bone mass. Although the exact threshold femur. Osteoporosis Int. 1, 147±154. of calcium intake required is uncertain, calcium supplemen- Cumming RG (1990): Calcium intake and bone mass: a quantitative review of the evidence. Calcif. Tissue Int. 47, 194. tation has been found to be effective in populations whose Dawson-Hughes B, Dallel GE, Krall EA, Sadoushi L, Sahyoner N & dietary calcium intake are below 400 mg (Dawson-Hughes Tannenbaum S (1990): A controlled trial of the effect of calcium et al, 1990). We have previously demonstrated that a low supplementation on bone density in postmenopausal women. New dietary calcium intake was associated with increased risk of Engl. J. Med. 323 (13), 878±883. Ellis FR, Holesh S & Ellis JW (1972): Incidence of osteoporosis in both hip fracture (Lau et al, 1988) and vertebral fracture vegetarians and omnivores. Am. J. Clin. Nutr. 25, 555±558. (Chan et al, 1996) in elderly Chinese women. Although the Food and Agriculture Organization (1972): Food Composition table for mean dietary calcium intake was found to be higher in use in South East Asia. Rome: FAO. Chinese vegetarians than omnivores in this study, the aver- Geinez G, Rapin CH, Rizzoli R, Kraemer R, Buchs B, Slosman D, Michel age dietary intake of 328 mg per day in vegetarians were still JP & Bonjour JP (1993): Relationship between bone mineral density and dietary intakes in elderly. Osteoporosis Int. 3, 242±248. low. The threshold required for protecting bone loss might Hegsted M & Linksweiler HM (1981): Long term effects of level of not have been reached in this group. protein intake or calcium metabolism in young adult women. J. Nutr. Alternatively, the effects of a high calcium intake 111, 244±251. among the vegetarians might have been off-set by high Hunt IF, Murphy NJ & Hendesson C et al (1989): Bone mineral content in postmenopausal women: comparison of omnivores and vegetarians. Am. levels of sodium consumption. In our study, tthe urinary J. Clin. Nutr. 50, 517±523. sodium to creatinine ratio in Chinese vegetarians was Kerstetter JE & Allen LH (1990): Dietary protein increases urinary almost twice as high as in the omnivores. It is known calcium. J. Nutr. 120, 134±136. that an increase in dietary sodium results in an increase in Lau EMC, Donnan SPB, Barker DJ & Cooper C (1988). Calcium intake urinary calcium excretion (Massey et al, 1968). On the and physical activity in fractured proximal femur in Hong Kong. Br. Med. J. 297, 1441±1443. average, urinary calcium excretion increased by around Lau EMC, Cooper C, Wickham C, Donnan S & Barker DJP (1990): Hip 0.6 mmol per 100 mmol increase in dietary sodium (Zar- fracture in Hong Kong and Britain. Int. J. Epid. 9(4), 1119±1121. kadas et al, 1989; Shortt et al, 1988). McParland et al Lloyd T, Schaffer JM, Walker MA & Demers M (1991): Urinary (1989) studied the biochemical effects of 10 d of low hormonal concentrations and spinal bone densities of premenopausal vegetarian and non-vegetarian women. Am. J. Clin. Nutr. 54, 1005± (70 mmol) and high (170 mmol=d) sodium intake. Serum 1010. osteocalcin increased by 8% and urinary hydroxyproline by Marsh AG, Sanchez TV, Mickelsen O, Keiser J & Mayor G (1980): 37% while subjects were put on a high salt diet. Although Cortical bone density of adult lacto-ovo-vegetarian and omnivorous the exact effect size of a high sodium intake on bone women. J. Am. Diet Assoc. 76, 148±152. density in human remains uncertain, its detrimental effect Marsh AG, Sanchez TW, Chaffee FL, Mayor GH & Mickelsen O (1983): Bone mineral mass in adult lacto-ovo-vegetarian and omnivorous male. on calcium metabolism and bone density is plausible. Am. J. Clin. Nutr. 37, 453±456. Marsh AG, Sanchez TV, Michelsen O, Chaffe FL & Fagal SM (1988): Vegetarian lifestyle and bone mineral density. Am. J. Clin. Nutr. 48, Conclusions 837±841. Massey SG, Coburn JW, Chapman LW & Kleeman CR (1968): Role of In our study, the BMD of `lactovegetarians' was not found serum Ca, parathyroid hormone, and NaCl infusion on renal Ca and Na to be signi®cantly different from `vegans'. Although the clearances. Am. J. Physiol. 214(6), 1403±1409. Bone mineral density in Chinese elderly females EMC Lau et al 64 McParland BE, Goulding A & Campbell AJ (1989): Dietary salt affects Shortt C, Madden A, Flynn A & Morrissey PA (1988): In¯uence of dietary biochemical markers of resorption and formation of bone in elderly sodium intake on urinary calcium excretion in selected Irish Individuals. women. Br. Med. J. 299, 834±835. Eur. J. Clin. Nutr. 42, 595±603. National Research Council (1980): Recommended Dietary Allowance. Tesar R, Notelovitzm, Shim E, Kauwell G & Brown J (1992): Axial and Committee on Dietary Allowance. Food and Nutrition Board, Division peripheral bone density and nutrient intake of post-menopausal vege- of Biological Sciences, Assembly of Life Sciences, Washington DC. tarian and omnivorous women. Am. J. Clin. Nutr. 56, 699±704. National Academy of Sciences. Tkatch L, Rapin CH, Rizzoli R, Slosman D, Nydegger V, Vasey H, Prince RL, Smith M, Dick IM, Price RI, Webb PT, Henderson NK & Bonjour JP (1992): Bene®ts of oral protein supplement in elderly Harris MM (1991): Prevention of postmenopausal osteoporosis. New patients with fracture of the proximal femur. Engl. J. Med. 325(17), 1189±1195. J. Am. Coll. Nutr. 11, 519±525. Pun KK, Chan LWL & Chung V (1989): The problem of Tylavsky FA & Anderson JJB (1988): Dietary factors in bone health of calcium de®ciency in Hong Kong. Hong Kong Practitioner 11, 287± elderly lacto-ovo-vegetarian and omnivorous women. Am. J. Clin. Nutr. 292. 48, 842±849. Schuette SA & Linkswiler HM (1982): Effects on Ca and P metabolism in Zarkadas M, Gougeon-Reyburn R, Marliss EB, Block E & Alton-Mackey humans by adding meat, meat plus milk, or puri®ed protein plus Ca and M (1989): Sodium chloride supplementation and urinary calcium P to a low protein diet. J. Nutr. 112, 338±349. excretion in post menopausal women. Am. J. Clin. Nutr. 50, 1088±1094.