Nutrient Intake and Iron Status of Australian Male Vegetarians

Nutrient intake and iron status of Australian male vegetarians

AK Wilson1 and MJ Ball1*

School of Nutrition and Public Health, Deakin University, Melbourne, Australia

Objective: The study was designed to investigate the iron intake and status of Australian, male vegetarians aged between 20 and 50 y. Design: Cross-sectional comparison of male vegetarians and age=sex matched omnivores. Setting: Free-living community subjects. Subjects: 39 ovolactovegetarians, 10 vegans and 25 omnivores were recruited by local advertisement. Outcome measures: A 12-d semiquantitative dietary record to assess iron and zinc intake. Iron status was assessed by measurement of serum ferritin and haemoglobin concentrations. Results: Mean (s.d.) daily iron intakes of both the ovolactovegetarians (20.4 (7.7) mg=d) and vegans (22.9 (6.2) mg=d), were signi®cantly higher than the omnivores' intake of 15.8 (4.5) mg=d. Ovo-lactovegetarians and vegans had signi®cantly (P < 0.001 and P < 0.05, respectively) lower serum ferritin concentrations than omnivores: mean (s.d.): 64 (46.9), 65 (49.9) and 121 (72.5) ng=ml, respectively. Signi®cantly more ovolactovegetarians and vegans than omnivores had serum ferritin concentrations below 25 ng=ml and below 12 ng=ml (P < 0.05). A higher proportion of omnivores had concentrations above 200 ng=ml (P < 0.05). The differences in serum ferritin concentrations between the vegetarians and omnivores remained signi®cant even after exclusion of iron supplement users. Conclusion: Australian male vegetarians had iron intakes higher than those of omnivores and above recommended levels, but their iron status was signi®cantly lower. Sponsorship: Deakin University. Descriptors: dietary iron; ferritin; iron status; vegetarianism

Introduction the only form of iron in vegetarian diets and is less bioavailable. In addition, food components that may limit the bio- Compared to omnivores, vegetarians have been reported to availability of non-haem iron are typically present in high have a lower incidence of a number of chronic diseases, amounts in vegetarian diets. These components include such as coronary heart disease (Havala & Dwyer, 1993), phytates (Oberleas & Harland, 1981; Hallberg et al, 1987; diabetes, osteoporosis, arthritis, kidney stones and some Ferguson et al. 1989), phenolic compounds (Brune et al, cancers (Dwyer, 1988). In addition, vegetarians generally 1989), calcium (Freeland Graves et al, 1980; Gibson, 1990) have lower blood pressure and body weight closer to and ®bre (Freeland Graves et al, 1980). desired levels (Dwyer, 1988). These factors may be attrib- The Australian Recommended Daily Intake (RDI) for iron uted to several dietary differences between omnivores and is 12 ± 15 mg for adolescents and menstruating women, and vegetarians and also to many positive lifestyle practices of 7 mg for men and post-menopausal women (NHMRC, 1991). vegetarians. As demonstrated by Burr and Butland (1988) it However, owing to the variation in iron bioavailability of is dif®cult to determine which aspects of the vegetarian diet different foods and in diets with various food combinations and lifestyle are protective. However, despite these health (Winston, 1994), assessment of adequacy of iron intake in bene®ts, the vegetarian diet has been associated with terms of meeting RDIs is a dif®cult task and subject to vitamin B , ribo¯avin, vitamin D, calcium, zinc and iron 12 considerable error. The daily zinc requirement for adults has de®ciencies (Gibson, 1990; Dwyer, 1991; Havala & been set in the range of 6 ± 12 mg (NHMRC, 1991). In the Dwyer, 1993). Australian National Nutrition Survey (McLennan & Podger, Owing to the physiological importance of iron and zinc, 1995), the mean daily intake for adult men in Victoria the adequacy of vegetarian diets with respect to both (determined from diet recall and food frequency question- minerals has been the focus of study by several investiga- naires (FFQ)) was 16.8 mg iron and 14.4 mg zinc. tors (Latta & Liebman, 1984; Levin et al, 1986; Alexander Several studies have been carried out to assess the et al, 1994; Donovan & Gibson, 1996). adequacy of vegetarian iron intake. However, few have Dietary iron intake is a concern in vegetarians, as animal used diet records over a 12 to 14-d period, considered foods are the exclusive source of haem iron, which is the most necessary to reliably determine iron intake (Basiotis et al, bioavailable form of dietary iron with an absorption ef®ciency 1987). One such study, conducted by Alexander and of 15 ± 25% (Monsen et al, 1978; Winston, 1994). Non-haem colleagues in New Zealand, found the iron intake of both iron, which is found in fruit, vegetables, cereals, dairy male and female vegetarians to be signi®cantly higher than products and eggs (Monsen et al, 1988; Winston, 1994), is that of omnivores (Alexander et al, 1994). This difference was particularly evident in males. These ®ndings are consistent with studies of Kelsay et al (1988) and others *Correspondence: M Ball, School of Nutrition and Public Health, Deakin (Levin et al, 1986; McEndree et al, 1983). University, Melbourne, Australia. Received 22 August 1998; revised 28 September 1998; accepted 12 Studies by Alexander et al (1994) and others (Faber et al, October 1998. 1986; Helman & Darnton Hill, 1987) assessed iron status by Nutrient and iron status of vegetarians AK Wilson and MJ Ball 190 measurement of the serum ferritin concentration, which is a sequences of 4 consecutive days followed by a few days measure of body iron status (Pilon et al, 1981). Results rest but overall including 3 ± 4 weekend days. The impor- showed both male and female vegetarians to have signi®- tance of the 12 d recording period being representative of cantly lower serum ferritin levels than omnivores. In other their normal eating patterns was explained to subjects. A studies, a signi®cant difference in iron status between vege- set of electronic scales was provided for each subject to tarians and omnivores was not shown when different methods weigh food. In situations where weighing of food was not of iron status assessment were used that are considered to have possible, subjects were instructed to estimate food portion low sensitivity relative to the serum ferritin measurement sizes by referring to photographs in the food dairy. (Latta & Liebman, 1984; Levin et al, 1986). After completion of the diary, it was examined to clarify A number of studies (Freeland Graves et al, 1980; Levin et any queries. A fasting, venous blood sample was taken for al, 1986) have also assessed both the adequacy of vegetarian estimation of serum ferritin and haemoglobin concentra- zinc intake and zinc status. Using a 24 h recall and FFQ as the tion. Weight and height were measured on a set of `Seca' diet assessment methods, results indicate there are no differ- digital scales, with minimal clothing and no shoes, and ences in vegetarian and omnivore zinc intake. However, a body mass index (BMI: weight (kg)=height (m2)) was study of adolescents by Donovan and Gibson (1996) using 3 d calculated. Waist and hip circumference were measured weighed records shows a higher probability estimate of to the nearest millimetre using a standard, non-extensible inadequacy for zinc and iron. Kadrabova et al (1995) also measuring tape, and the waist=hip ratio was calculated. found plasma zinc concentration in vegetarians to be signi®- Sitting systolic and diastolic blood pressure were deter- cantly lower than in omnivores, although they provided no mined using an oscillomate 93 (Wallace Instruments Pty dietary data and plasma zinc may not be a particularly reliable Ltd, Australia). measure of zinc status. Interestingly, the small study by Alexander et al (1994) Haematological measurements raised particular concern for the iron status of vegetarian Haemoglobin was measured immediately using an electro- men, as a signi®cant proportion had serum ferritin levels nic counter (HemoCue AB: Mallinckrodt Medical Pty Ltd, below 12 ng=ml, indicating storage iron depletion. Despite Nottinghill, Australia). Serum was separated by centrifuga- these results, and the increasing incidence of male vegetar- tion at 4C for 15 min (3000 rpm) after allowing the blood ianism in Australia, most concern is typically given to to clot at room temperature for 1 h. Serum samples were women, with very little focus on men. In fact there has been stored at 7 20C for later analysis of serum ferritin. Serum only one published Australian study investigating iron ferritin measurements were performed at the `Inter-Lab' status of vegetarian males (Helman & Darnton Hill, pathology laboratory by a two-site chemiluminometric 1987), and none speci®cally addressing the adequacy of immunoassay, using a Ciba Corning Automated Chemilu- iron and zinc intake. minescence System (ACS: 180 Australian Diagnostics The aim of this study was to assess the iron and zinc Corporation Pty Ltd, Ciba Corning Diagnostics Corp, intake and iron status of healthy Australian male adult Scoresby). vegetarians using a 12 d semiquantitative food record and serum ferritin and haemoglobin measurements and to Diet analysis compare the results with age- and sex-matched omnivores. The food and beverage intakes recorded on the 12 days by each participant were entered into the dietary analysis Methods program `Diet-3', whose data is based on Australian Food Composition Tables (Xyris 1987 ± 91, version 3.22). When Fifty vegetarians aged 20 ± 50 y were recruited in the composite foods eaten were not present on the database, Melbourne Metropolitan area, by advertisement in maga- recipes were entered. zines, restaurants and health food shops. Twenty-®ve healthy omnivores, of the same age and similar body Statistical analysis mass index (BMI) to the vegetarian subjects, were also recruited as `friend controls' of the vegetarians and by All statistical analyses were performed using the Statistical advertisement. For the purpose of this study, an ovolacto- Package for Social Sciences version 6.0 (SPSS Pty Ltd, vegetarian was de®ned as someone who never consumed Chicago, IL, USA). Nonparametric analyses using the Krus- red meat and consumed chicken or ®sh no more than once a kal ± Wallis one-way ANOVA were performed to determine week; a vegan was de®ned as someone who ate no meat, differences between omnivores, ovolactovegetarians and eggs or dairy products. Unless otherwise speci®ed, `vege- vegans for nutrient intake and haematological measurements. tarian' includes both of these groups. An omnivore was The Mann ± Whitney U-test was then performed with the someone who consumed red meat. Subjects were included Bonferroni adjustment (Matthews & Farewell, 1988) to only if they had been practising the particular diet for at identify which groups had signi®cant differences. Multiple least 6 months, were non-smokers, were in good health and regression analyses were performed to assess the relationship were not using any medications regularly. Deakin Univer- between dietary variables and serum ferritin concentrations. sity Ethics Committee approved this study, and each The Fisher exact test was used to demonstrate whether the participant gave written informed consent. three dietary groups were different with respect to having At an initial meeting, all subjects completed a ques- values below and above certain limits. tionnaire on their dietary habits, use of supplements and medications, previous de®ciencies related to diet, smoking Results habit and usual levels of exercise. Subjects were given detailed instructions on how to complete accurately a The vegetarian group that completed the study consisted of validated, 12 d semiquantitative food record (Edington et 39 ovolactovegetarians and 10 vegans. The ovolactovege- al, 1988) on either 12 consecutive days or from three tarians and vegans had been following their dietary Nutrient and iron status of vegetarians AK Wilson and MJ Ball 191 practices for a mean of 11 y (range 0.5 ± 44) and 6 y (range The mean daily iron and zinc intakes of omnivores, 1.5 ± 14), respectively. The 25 omnivore controls who ovolactovegatarians and vegans are shown in Table 3. Ovo- participated did not differ signi®cantly from either the lactovegetarians and vegans had signi®cantly higher mean ovolactovegetarians or vegans with respect to age, daily iron intakes than omnivores (P < 0.05 and P < 0.01, BMI, waist=hip ratio, and blood pressure (Table 1), respectively). All the men had an intake above the RDI for although their weight was higher than the ovolacto- men of 7 mg=d, the mean intakes for omnivores, ovolactovegatarians. vegetarians and vegans being 226%, 291% and 320% of The daily nutrient intakes of the three dietary groups are RDI, respectively. Zinc intakes of the three groups were not shown in Table 2. While total energy intake was similar for signi®cantly different, with a high proportion of the omni- omnivores, ovolactovegetarians and vegans, the vegetar- vores, ovolactovegetarians and vegans having intakes ians derived a signi®cantly lower proportion of their energy below the Australian RDI of 12 mg=d (40%, 64% and from protein compared to omnivores. A much higher 70%, respectively). proportion of energy derived from carbohydrate compen- The dietary sources of iron of the omnivores and sated for this. Total fat intake and percentage of energy vegetarians are indicated in Figure 1. Both ovolactovege- derived from fat were lower for ovolactovegetarians than tarians and vegans obtained their iron predominantly from for omnivores, as was saturated fat intake, with vegans cereals, vegetables, fruits, seeds and nuts. Cereals were also consuming less than half the saturated fat of omnivores. a main source of dietary iron for omnivores, but, in contrast The vegetarians consumed a higher percentage of total fat to vegetarians, omnivores obtained a high proportion of as polyunsaturated fat. Cholesterol intake was signi®cantly iron from meat and meat products. different between all three dietary groups, with the omni- Serum ferritin and haemoglobin concentrations (Table vores consuming the most and the vegans the least. Dietary 4) were signi®cantly lower in ovolactovegetarians and ®bre was signi®cantly higher in vegetarians than in omni- vegans compared to those in omnivores (P < 0.001 and vores. Mean calcium intakes did not differ signi®cantly and P < 0.05, respectively). The distribution of serum ferritin were above the Australian RDI of 800 mg=d. Vitamin C levels is shown in Figure 2. Differences between the three intake of all groups was also well above recommended dietary groups were signi®cant, with more ovolactovege- levels, with the vegans consuming signi®cantly greater tarians and vegans than omnivores having serum ferritin amounts, with a mean of nine times the Australian RDI concentrations below 25 ng=ml (P < 0.05). One ovolacto- of 40 mg=d (NHMRC, 1991). vegetarian aged 23 years and two vegans aged 41 and 36

Table 1 Physical characteristics of omnivores, ovolactovegetarians and vegans; mean (s.d.)

Age Weight Height Systolic Diastolic Group n (y) (kg) (cm) BMI BP BP

Omnivores 25 32.7 79.8 179.9 23.8 124.8 75.8 (8.8) (9.5) (5.9) (2.7) (15.7) (8.4) Ovolactovegetarians 39 33.3 70.6* 179.5 21.8 115.3 72.4 (8.2) (11.2) (7.9) (2.5) (12.6) (12.8) Vegans 10 31.0 74.1 180.5 22.6 119.3 73.5 (5.6) (14.2) (8.0) (3.1) (9.5) (9.6)

*Signi®cantly different from omnivores, P < 0.001.

Table 2 Comparison of daily nutrient intakes of male ovolactovegetarians vegans and omnivores; mean values or mean (s.d.)

Ovo-lactovegetarians Vegans Nutrient Omnivore (n=25) (n=39) (n=10)

Energy (MJ) 11.0 (1.9) 10.5 (2.4) 11.6 (2.7) [Kcal] [2632] [2548] [2815] Protein (g) 108 (17.2) 80 (16.4)b 81 (24.3)a % total energy 17.3 15.5b 12.4b Fat (g) 98.3 (31.8) 81.8 (28.1)a 88 (37.6) % total energy 33.5 28.2a 28.0 Carbohydrate (g) 291 (77.4) 357 (92) 413.4 (96.8) % total energy 44.9 54.5b 59.1b Polyunsaturated fat (g) 14.6 (8.4) 19.7 (8.6) 27.2 (16.0) % total fat 16.1 27.5b 31.8b Monounsaturated fat (g) 34.6 (11.7) 28.9 (15.0)a 35.4 (16.3) % total fat 38.1 38.3c 44.1a Saturated fat (g) 41.3 (13.3) 24.2 (9.1)b 18.3 (7.5)b %total fat 45.8 35.0bd 24.1b Cholesterol (mg) 289 (96) 102 (67)bc 22 (28)b Dietary ®bre (g) 26.3 (7.7) 49.8 (15.9)b 63.8 (14.6)b Sodium (g) 3.4 (0.7) 2.5 (0.8)b 1.7 (9.5)b Calcium (mg) 961 (263) 899 (284) 911 (471) Vitamin C (mg) 151 (93) 218 (178)c 360 (179)a Vitamin A (mg) 981 (326) 1266 (602) 1661 (497)b a,bSigni®cantly different from omnivores (aP < 0.01, bP < 0.001). c,d,eSigni®cantly different between ovolactovegetarians and vegans (cP < 0.05, P < 0.01, eP < 0.001). Nutrient and iron status of vegetarians AK Wilson and MJ Ball 192 Table 3 Daily intakes of iron and zinc of omnivores, ovolactovegetarians and vegans. Mean intakes for iron are given as a percentage of Australian RDI. The proportion of subjects with zinc intakes below the Australian RDI is also given

Iron intake (mg=d) Zinc intake (mg=d)

Group n Mean (s.d.) Range % of RDI Mean (s.d.) Range < RDI (%)b

Omnivores 25 15.8 (4.5) 8.2 ± 27.5 226 12.7 (2.4) 9.3 ± 16.9 40 Ovo-lactovegetarians 39 20.4 (7.7a 8.5 ± 42.3 291 12.0 (4.4) 5.8 ± 26.1 64 Vegans 10 22.9 (6.2)b 15.2 ± 35.1 320 12.6 (3.8) 7.4 ± 18.1 70

aThe mean iron intake as a percentage of the Australian RDI for iron of 7 mg=d. bThe percentage of subjects below the Australian RDI for zinc of 12 mg=d. a,bSigni®cantly different from omnivores (aP < 0.05, bP < 0.01).

ferritin concentration of 153 ng=ml (in the normal range), while the other two subjects had low concentrations. Eight ovolactovegetarians and two omnivores were either currently taking iron supplements or had taken them in the previous two years. The serum ferritin concentrations of this group ranged from 24 to 117 ng=ml. Twenty-one vegetarians (including 4 who also took iron supplements) currently took a vitamin supplement that did not contain iron. Additionally, 11 ovolactovegetarians consumed either ®sh a few times a year or chicken once or twice a month (three of whom also took iron supplements). After exclusion of all iron-supplement users, and of ovolactovegetarian ®sh=chicken eaters the serum ferritin levels of omnivores remained signi®cantly higher than those of the ovolactovegetarians (P < 0.001) and vegans (P < 0.01). Multiple regression analysis showed there was no sig- Figure 1 Serum ferritin concentration (ng=ml) in vegetarians and omni- ni®cant correlation between age, alcohol, calcium, iron or vores. ®bre intake and serum ferritin concentrations for the vegetarians and omnivores. However, a highly signi®cant, positive correlation was found between vitamin C intake years had serum ferritin concentrations below 12 ng=ml and serum ferritin concentrations (r 2 0.52, P < 0.001). (11, 8 and 10 ng=ml, respectively). All three subjects had iron and vitamin C intakes well above Australian RDIs and had not taken iron supplements, and the ovolactovegetarian Discussion consumed ®sh and chicken once or twice a month. More omnivores than vegetarians had serum ferritin concentra- The adequacy of the vegetarian diet with respect to nutrient tions above 200 ng=ml (P < 0.05). and vitamin intake has been the focus of a number of Three subjects, two ovolactovegetarians and one vegan, investigations. However, the present study is one of only had haemoglobin concentrations (125, 129 and 120 g=l) two Australian studies that we are aware of that have below the reference range of 130 ± 150 g=l: one had a serum assessed the adequacy of adult male vegetarian iron

Figure 2 Dietary sources of iron in omnivores (j) and vegetarians (u). Nutrient and iron status of vegetarians AK Wilson and MJ Ball 193 Table 4 Subjects with serum ferritin in certain ranges and haemoglobin (g=L) concentrations in omnivores, ovo-lactovegetarians and vegans

HB (g=l) Mean Ferritin Ferritin Ferritin Group n (s.d.) Range < 12 ng=ml n (%)a < 25 ng=ml n (%)a > 200 ng=ml n (%)a Mean (s.d.) Range

Omnivores 25 121 (73) 44 ± 313 0 (0) 0 (0) 5 (20) 173 (19) 144 ± 206 Ovo-lactovegetarians 39 64 (47)c 11 ± 238 1 (3) 8 (20.5)b 1 (2.6)b 140 (40)c 125 ± 195 Vegans 10 65 (50)b 8 ± 137 2 (25) 3 (30) 0 (0)b 158 (28)b 129 ± 201 aNumber and percentage of subjects with serum ferritin in the stated range. c,dSigni®cantly different from omnivores (bP < 0.05, cP < 0.001). status. Adequacy of iron intake was assessed using a prove insuf®cient to provide for periods of stress or an quantitative food record over an adequate period of time, episode of blood loss, and continued low availability of and serum ferritin and haemoglobin concentrations were iron may result in iron de®ciency. The fact that iron intakes measured as an indication of iron status. of all subjects with ferritin levels below 25 ng=ml were Iron intakes of both ovolactovegetarians and vegans similar to or higher than those of other vegetarians and were signi®cantly higher than those of omnivores, with omnivores with ferritin concentrations in the normal range, all dietary groups having intakes well above the Australian indicates inter-person variation in absorption. Intake of RDI of 7 mg=d. A study conducted by Alexander et al ascorbic acid was well above recommended intakes for (1994) on a vegetarian population in New Zealand consist- these individuals, and this is relevant as absorption of non- ing of 36 females and 14 males also found male vegetarian haem iron is enhanced by consumption of ascorbic acid in iron intake to be above recommended levels and higher the same meal (Cook & Monsen, 1977). However, overall than that of omnivores. The intakes (mean 20.4 mg=d and ef®ciency of non-haem iron absorption is dependent on a 20.2 mg=d, respectively), using the same dietary methodol- balance between inhibiting components and ascorbic acid ogy, were very similar to those in our study. The differ- intake (Hallberg & Rossander, 1982). Vegetarians whose ences in iron intake between vegetarians and omnivores are non-haem iron and ascorbic acid intakes are similar may consistent with those found by Kelsay et al (1988), who have different levels of serum ferritin. Young Chinese compared iron intakes of Asian-Indian and American Buddhist vegetarians have half the ferritin level of nonvegetarian groups to an American omnivore group. Other vegetarian fellow students, and this was at least partly investigators ®nding similar results include Levin et al attributed to their high phytate and soy intake (Shaw et (1986), who studied an Israeli population consisting of 92 al, 1995). The higher serum ferritin levels in omnivores can vegetarians and 113 omnivores, and McEndree et al (1983), easily be explained by the consumption of the more who investigated a vegetarian group in Nebraska. ef®ciently absorbed haem iron. A common methodological feature of our study and the Interestingly, one ovolactovegetarian who ate chicken studies conducted by Alexander et al (1994) and Kelsay et and ®sh a few times a year, and 20% of omnivores, had al (1988) is that nutrient intakes were assessed using 12 d serum ferritin concentrations above 300 ng=ml. Such high and 14 d semiquantitative, diet records. This method of levels may not be advantageous. In the Kuopio Ischaemic dietary assessment has been shown by Edington et al Heart Disease Risk Factor Study (Salonen, 1993), a mean (1988) to have similar reliability to weighed food records, serum ferritin concentration greater than 200 ng=ml was which are considered the most precise method. As calcu- found to correspond to a 2.2-fold increased risk of acute lated by Bingham (1987), 12 days are required to record myocardial infarction. Although these ®ndings have not individual, long-term iron intake within 10% of the mean. been con®rmed in subsequent American studies (and dis- In contrast, studies that used records over shorter periods of agreement in results could be explained by differences in 3 ± 7 days (Latta & Liebman, 1984; Faber et al, 1986) have serum cholesterol levels, as the risk due to elevated serum not shown a difference in vegetarian and omnivore iron ferritin decreases when cholesterol levels are below intake. 5 mmol=l), it may indicate relative advantage for most Ovolactovegetarians obtained their dietary iron exclu- vegetarians. sively from non-haem iron sources such as fruits, vegeta- Signi®cant differences in iron status between vegetar- bles, cereals, dairy products and eggs, while for vegans, ians and omnivores are consistent with ®ndings by Alex- fruit, vegetables and cereals were the only iron source. ander et al (1994) and others (Faber et al, 21986; Helman Although the amount of iron consumed was above the RDI, & Darnton Hill, 1987). Serum ferritin concentrations were the actual amount absorbed may be considerably lower, lower for both vegetarians and omnivores in the New owing to the low bioavailability of non-haem iron and food Zealand population compared to those in the Australian components inhibiting absorption. study, despite similar iron intakes (Alexander et al, 1994). Iron status was thus measured and, as indicated by This may be explained by the fact the Australian vegetar- serum ferritin concentration, was signi®cantly lower in ians consumed more vitamin C than the New Zealanders. the vegetarians compared to the omnivores. Twenty-eight Helman and colleagues (Helman & Darnton Hill, 1987) per cent of vegetarians had serum ferritin levels below also studied an Australian vegetarian population (60 males 25 ng=ml, which may indicate the beginning of iron storage and 60 females) but did not consider males and females depletion (Bothwell et al, 1979). Furthermore, one ovolac- separately, so comparison is not possible. There are, how- tovegetarian and two vegans had serum ferritin concentra- ever, a number of papers reporting no signi®cant difference tions below 12 ng=ml, indicating depletion of storage iron in vegetarian and omnivore iron status. In some cases this (Cook et al, 1992). Although these subjects' haemoglobin may be due to use of supplements by many subjects; for levels were in the normal range, these storage levels might example, in the study by Locong (1986), 50% of the Nutrient and iron status of vegetarians AK Wilson and MJ Ball 194 vegetarians studied (22 females and 14 males) ate ®sh= Donovan UM & Gibson RS (1996): Dietary intakes of adolescent females poultry once a month and 61% took vitamin and mineral consumingvegetarian,semi-vegetarian, and omnivorousdiets. J. Adolesc. Health 18, 292 ± 300. supplements. Animal ¯esh enhances iron absorption by Dwyer JT (1988): Health aspects of vegetarian diets. Am. J. Clin. Nutr. 48, stimulating the absorption of non-haem iron and providing 712 ± 738. some haem iron (Hallberg et al, 1986). In other studies, the Dwyer JT (1991): Nutritional consequences of vegetarianism. Annu. Rev. method of iron status assessment used is less sensitive than Nutr. 11, 61 ± 91. serum ferritin (Pilon et al, 1981). Latta and Leibman (1984) Edington J, Thorogood M, Geekie M, Ball M & Mann J (1988): Assess- ment of nutritional intake using dietary records with estimated weights. assessed the iron status of 36 American male vegetarians by J. Hum. Nutr. Diet. 2, 407 ± 414. determination of whole-blood hematocrit, plasma iron and Faber M, Gouws E, Benade AJ & Labadarios D (1986): Anthropometric total iron binding capacity, and there can be poor iron status measurements, dietary intake and biochemical data of South African when these parameters are normal. lacto-ovovegetarians. S. Afr. Med. J. 69, 733 ± 738. Ferguson EL, Gibson RS, Thompson LU & Ounpuu S (1989): Dietary As in previous studies (Freeland Graves et al, 1980; calcium, phytate, and zinc intakes and the calcium, phytate, and zinc Levin et al, 1980; Levin et al, 1986), zinc intakes in the molar ratios of the diets of a selected group of East African children. ovolactovegetarian, vegans and omnivores were found to Am. J. Clin. Nutr. 50, 1450 ± 1456. be similar. Dietary phytates (Oberleas & Harland, 1981), Freeland Graves JH, Bodzy PW & Eppright MA (1980): Zinc status of calcium and insoluble ®bre (Freeland Graves et al, 1980) vegetarians. J. Am. Diet. Assoc. 77, 655 ± 61. Gibson R (1990): Principles of Nutritional Assessment, Oxford University inhibit zinc absorption, and the high content of these food Press, New York. components and the exclusion of animal meats from the Hallberg L & Rossander L (1982): Effect of different drinks on the vegetarian diet may lead to a low zinc status in vegetarians. absorption of non-heme iron from composite meals. Hum. Nutr. Appl. In the present study, a high proportion of subjects in all Nutr. 36, 116 ± 123. three dietary groups had zinc intakes below the Australian Hallberg L, Brune M & Rossander L (1986): Effect of ascorbic acid on iron absorption from different types of meals. Studies with ascorbic- RDI of 12 mg=d. Therefore, consideration should be given acid-rich foods and synthetic ascorbic acid given in different amounts to measurement of zinc status in vegetarians who consume with different meals. Hum. Nutr. Appl. Nutr. 40, 97 ± 113. large amounts of calcium and ®bre and phytate-rich foods Hallberg L, Rossander L & Skanberg AB (1987): Phytates and the such as rye bread, apples, rice, oatmeal and wheat cereals inhibitory effect of bran on iron absorption in man. Am. J. Clin. Nutr. 45, 988 ± 996. (Turnlund et al, 1984). Havala S & Dwyer J (1993): Position of the American Dietetic Associa- The present study demonstrates that vegetarians had tion: vegetarian diets [published erratum appears in J. Am. Diet. Assoc. intakes of most macronutrients, minerals and vitamins 1994 94(1), 19] J. Am. Diet. Assoc. 93, 1317 ± 1319. closer to Recommended Intakes than did omnivores, parti- Helman AD & Darnton Hill I (1987): Vitamin and iron status in new cularly with respect to quantity and type of fat intake, vegetarians. Am. J. Clin. Nutr. 45, 785 ± 789. Kadrabova J, Madaric A, Kovacikova Z & Ginter E. (1995): Selenium sodium, antioxidant vitamin and ®bre intake. This may status, plasma zinc, copper, and magnesium in vegetarians. Biol. Trace provide health bene®ts, particularly with respect to coron- Elem. Res. 50, 13 ± 24. ary heart disease. However, more consideration needs to be Kelsay JL, Frazier CW, Prather ES, Canary JJ, Clark WM & Powell AS given to iron status in these men, even when their intake (1988): Impact of variation in carbohydrate intake on mineral utilization by vegetarians. Am. J. Clin. Nutr. 48, 875 ± 879. appears to meet the RDI. In fact, the RDI may need to be Latta D & Liebman M (1984): Iron and zinc status of vegetarian and higher than 7 mg for male vegetarians. Attention should be nonvegetarian males. Nutr. Rep. Int. 30, 141 ± 149. given to optimising non-haem iron absorption by education Levin N, Rattan J & Gilat T (1986): Mineral intake and blood levels in about meal planning. Main meals containing non-haem iron vegetarians. Isr. J. Med. Sci. 22, 105 ± 108. should be eaten in combination with foods high in ascorbic Locong A (1986): Nutritional status and dietary intake of a selected sample of young adult vegetarians. J. Can. Diet. Assoc. 47, 101 ± 106. acid while high phytate, phenolic and calcium foods should Matthews D & Farewell V (1988): Using and Understanding Medical generally be avoided at the same time. Statistics. Karger Publications, Basel. McEndree L, Kies C & Fox H (1983): Iron intake and nutritional status of Acknowledgements ÐWe thank Professor Andy Sinclair, Duo Li and Neil lacto-ovo vegetarian and omnivore students eating in a lacto-ovo Mann, RMIT for assisting in recruiting volunteers. vegetarian food service. Nutr. Rep. Int. 27, 199 ± 206. McLennan W & Podger A (1995): National Nutrition SurveyÐSelected Highlights in Australia. Canberra: Australian Bureau of Statistics. References Monsen ER (1988): Iron nutrition and absorption: dietary factors which impact iron bioavailability. J. Am. Diet. Assoc. 88, 786 ± 790. Alexander D, Ball MJ & Mann J (1994): Nutrient intake and haematolo- Monsen ER, Hallberg L, Layrisse M, Hegsted DM, Cook JD, Mertz W & gical status of vegetarians and age-sex matched omnivores. Eur. J. Clin. Finch CA (1978): Estimation of available dietary iron. Am. J. Clin. Nutr. 48, 538 ± 546. Nutr. 31, 134 ± 141. Basiotis PP, Welsh SO, Cronin FJ, Kelsay JL & Mertz W. (1987): Number NHMRC (1991): NHMRCÐRecommended Dietary Intakes for Use in of days of food intake records required to estimate individual and group Australia. Australian Government Publishing Service. nutrient intakes with de®ned con®dence. J. Nutr. 117, 1638 ± 1641. Oberleas D & Harland BF (1981): Phytate content of foods: effect on Bingham S (1987): The dietary assessment of individuals; methods, dietary zinc bioavailability. J. Am. Diet. Assoc. 79, 433 ± 436. accuracy, new techniques and recommendations. Nutr. Abstr. Rev. 7, Pilon VA, Howanitz PJ, Howanitz JH & Domres N. (1981): Day-to d 706 ± 742 variation in serum ferritin concentration in healthy subjects. Clin. Chem. Bothwell TH, Charlton RW & Cook JD (1979): In: Iron Metabolism in 27, 78 ± 82. Man, TH Bothwell (ed). Blackwell Scienti®c: Oxford. Salonen JT (1993): The role of iron as a cardiovascular risk factor. Curr. Brune M, Rossander L & Hallberg L. (1989): Iron absorption and phenolic Opin. Lipidol. 4, 277 ± 282. compounds: importance of different phenolic structures. Eur. J. Clin. Shaw NS, Chin CJ & Pan WH (1995): A vegetarian diet rich in soybean Nutr. 43, 547 ± 557. products compromises iron status in young students. J. Nutr. 125, 212 ± Burr ML & Butland BK. (1988): Heart disease in British vegetarians. Am. 219. J. Clin. Nutr. 48, 830 ± 832. Turnlund JR, King JC, Keyes WR, Gong B & Michel MC (1984): A stable Cook JD & Monsen ER (1986): Vitamin C, the common cold, and iron isotope study of zinc absorption in young men: effects of phytate and absorption. Am. J. Clin. Nutr. 30, 235 ± 241. alpha-cellulose. Am. J. Clin. Nutr. 40, 1071 ± 1077. Cook JD, Baynes RD & Skiknes BS (1992): Iron de®ciency and the Winston CJ (1994): Iron status of vegetarians. Am. J. Clin. Nutr. 59, measurement of iron status. Nutr. Res. Rev. 5, 189 ± 202. 1233 ± 1236.