European Journal of Clinical (1997) 51, 839±845 ß 1997 Stockton Press. All rights reserved 0954±3007/97 $12.00 Quantitative responses of serum to increasing intakes of folic acid in healthy women

AS Truswell and S Kounnavong

Human Nutrition Unit, Department of G08, University of Sydney, NSW 2006,

Objectives: Health authorities are advising people to increase folate intake and more foods are being forti®ed with folic acid. Estimation of dietary folate is dif®cult because data bases lack this nutrient in many countries and is variable. To see if serum folate can be used to re¯ect effective folate intake, we have measured serum folate after healthy women had taken different doses of pure folic acid supplements in the nutritional range. Methods: A of 20 volunteer subjects took part in one or more of three experiments, six of them took part in all three. In each experiment subjects took two different doses of folic acid, each for three weeks. Experiment (1) 100 mg then 1000 mg=d; experiment (2) 500 mg then 1500 mg=d; experiment (3) 1000 then 2000 mg=d. Serum folate was measured after overnight fast with a homogenous binding assay kit. In experiment (3) red cell were also measured. Results: It took three weeks' of the same supplement for serum folate to reach its full higher value. Mean serum folates of small groups of subjects were signi®cantly higher with each higher dose of supplement. The biggest increase was for the ®rst 100 mg, that is in the range of intake from unforti®ed foods. The curve of serum folate against folic acid supplement intake shows that serum folate of groups of people, properly standardized, can provide a practical method for assessing adequacy or change of people's folate intake. Red cell folates were less responsive. Sponsorship: S Kounnavong was in receipt of an Australian Agency for International Development Fellowship. Expenses were covered by the University of Sydney and its Nutrition Research Foundation. There was no con¯ict of interest. Descriptors: Serum folate; assessment of nutrient intake; biomarker; folic acid supplements; red cell folate

Introduction tation) or by forti®cation of staple foods. Since many are unplanned and extra folate is needed only Since publication of the preventive effects of folic acid in the ®rst four weeks after conception, food forti®cation against neural tube defects in secondary (Medical Research should have a broader reach than nutrition education or Council, 1991) and primary prevention trials (Czeizel & supplementation (which is most suitable for secondary Dudas, 1992), supported by other epidemiological studies prevention). Some breakfast are now voluntarily (Bower, 1996) government health departments in several forti®ed with folic acid in Britain, the USA and Australia. countries recommend measures to increase folate intake in In the USA mandatory forti®cation of enriched , rolls, women who could become pregnant (Centers for Disease ¯ours, meal, and has been announced by Control, 1992; Expert Advisory Group, 1992; Health and the Food and Drug Administration (1996). Welfare , 1993; Anonymous, 1995; Staatstoezicht With authoritative bodies actively encouraging in- op de Volksgesondheid, 1993; National Health and Medical creased folate intake by three different routes, it becomes Research Council, 1995). Women and their babies are not important to be able to monitor the folate intake of samples the only groups that may bene®t from more generous folate of the population and how these change. We would like to intake. Mild to moderate homocysteinaemia is being recog- know whether intakes have increased and to what ranges, nized as one of the risk factors for cardiovascular disease which sections of the community have and which have not (Bouskey et al, 1995) and probably for this reason low responded, and whether any individuals are consuming serum folate levels are associated with increased risk of very high levels that might lead to nutritional imbalance. coronary heart disease (Morrison et al, 1996). It is also To estimate folate intakes is more dif®cult that estimat- possible that dietary folate gives some protection against ing, say, thiamin or intakes. Numerous folate colorectal (Giovanucci et al, 1993; Cravo et al, compounds make food analysis complicated and dif®cult 1994). to standardize (Scheelings, 1996). While a microbiological People's folate intake can be increased in three ways: by method (following deconjugation) can give an overall choosing more foods with high (natural) folate content value, individual folate compounds may have different (nutrition education), with folic acid tablets (supplemen- biological activity in (Phillips & Wright, 1982) or in humans, but only a minority of the individual can be quanti®ed at present by HPLC. In many Correspondence: Professor AS Truswell countries there are no values for folate content of local Received 7 February 1997; revised 23 July 1997; accepted 28 July 1997 foods. There is also a shortage of information about the Quantitative reponses of serum folate AS Truswell and S Kounnavong 840 Table 1 Subject and doses of folic acid supplements they took

Experiment No. of subjects Ages mean BMI mean Oral contraceptive Smokers Doses of (range) years (range) kg=m2 users folic acid mg

1 13 27(21±36) 23(19±36) 5 3 100 then 1000 2 16 27(21±36) 22(19±27) 5 2 500 then 1500 3 6 29(21±36) 22(19±27) 2 0 1000 then 2000

effects of different methods of food processing and pre- ments, namely they took six three-week courses of folic paration on the bioavailability of folate(s). acid supplements. It seemed to us, therefore, that a biomarker method might by-pass both the subjectivity of food intake methods and the incomplete information about folate content of Methods local foods. The situation is thus somewhat analogous to Healthy young non pregnant women were volunteers, estimating intake, for which 24 h urinary sodium (if recruited from the University community. Their health repeated or from suf®cient numbers of subjects) is a more and food habits were checked by questionnaire. Usual reliable and less laborious method than trying to work out folate intake was estimated by a simpli®ed food frequency the content of and other sodium compounds in different questionnaire emphasising sources of folate and also at the individual foods and the salt retention with different cook- start of the second experiment with a four-day food intake ing methods. diary. As Australian food tables do not include values for In the literature serum folate is said to re¯ect recent folate, British (Holland et al, 1991) and intake while red cell folate indicates tissue content. Serum (Burlingame et al, 1993) food composition tables were folate would therefore seem to have the better potential as used to estimate folate intakes. indicator of dietary intake. We have, however, been unable In each of these experiments subjects were asked to take to ®nd the answer to two questions: a supplement of pure folic acid at one dose for three weeks, (i) What is the time interval between a change of folate then at a higher dose for the next three weeks (Table 1). intake and of serum folate concentration? Sauberlich et al Venous bloods were taken (with subjects lying down) after (1974) noted that `unfortunately the number of subjects overnight fast and before breakfast on ®ve occasions: studied on controlled intakes of folacin has been very before starting folic acid, after two and three weeks on limited. Low serum levels may be a re¯ection of only the lower dose and after two and three weeks on the higher low dietary intakes of the and provide little dose. The two week bloods were omitted in experiment (3). information concerning tissue reserves.' Participants were instructed to maintain their usual diets According to Gibson (1990), `serum folate concentrations and lifestyle, to keep alcohol consumption low and report ¯uctuate rapidly with recent changes in folate intakes and any illness. with temporary changes in folate , even when Six subjects took part in all three experiments, so they body stores remain stable. For example serum folate values took folic acid supplements of 100, 500, 1000 mg (twice), increase rapidly after the ingestion of folate-containing 1500 and 2000 mg. Another three subjects took part in foods and supplements. . .and decrease rapidly on a experiments (1) and (2). There was an interval of 28 d folate-de®cient diet, stabilizing at values below 3 ng=ml without the subjects taking supplements between experi- after only two to three weeks' of negative folate balance.' ments (1) and (2) and experiments (2) and (3). Kohlmeier (1995) in a table of biomarkers shows serum The pure folic acid (Roche , French's Forest, folate as indicating dietary exposure for the `past few NSW Australia) was dissolved in 0.1 M aqueous sodium days'. bicarbonate solution and the concentration in each experi- (ii) The other question is what is the shape of (or equation ment was such that the daily dose was obtained from 5 ml for) the relationship between folate intake and serum twice a day, which the subjects took by syringe (without folate? It cannot be assumed that the relationship is needle) directly into the mouth. Solutions were made fresh quantitative and linear. Serum levels of , ascorbate in amber glass bottles and given to the subjects every 7 d. and calcium, for example, all have different relationships Subjects were instructed to keep the bottles in their domes- with recent intake of the nutrient, none of them linear. tic refrigerator. Under these conditions we found minimal loss of folate from the solution by laboratory testing. There is no report of serum folate in groups of the same Compliance was recorded by the subjects ticking date and healthy subjects at several different intake levels (with time of each dose on a reminder sheet, and by retrospec- adequate time for adjustment) using the same method and tively measuring the contents of the bottles brought back to standardization throughout. In preliminary experiments the laboratory each week. (Leece et al, 1995; Truswell, 1996) we gave 100 mg=d Serum was obtained by centrifugation after the pure folic acid twice daily to 17 healthy young women. samples had clotted, and stored at 7 80C until the end of Serum folates were measured every few days. Mean serum each experiment when all the sera were analysed in one folate reached 70% of its maximum increase by 10 d batch. but did not reach its maximum until between days 21 and Serum folate was analyzed by homogeneous enzyme 24. binding assay kit (CEDIA vitamin B-12 and Folate Assays, In the work reported here, three experiments were Microgenics Corporation, Concord CA, USA). The kit carried out, in each of which healthy women were given contains b-galactoside that had been split into two totally folic acid supplements at two different doses, each for three inactive fractions, enzyme acceptor (EA, 95% of the b- weeks. Six of the subjects participated in all three experi- galactoside) and enzyme donor (ED, 5% of the b-galacto- Quantitative responses of serum folate AS Truswell and S Kounnavonget al 841 Table 2 Serum folates (mg=ml) after two and three weeks of folic acid

Folic acid supplement ‡ 100 mg ‡ 500 mg ‡ 1000 mg ‡ 1500 mg

2 wks 3 wks 2 wks 3 wks 2 wks 3 wks 2 wks 3 wks

6.55 8.22 10.30 11.03 13.12 13.38 13.33 13.84 2 wks =3 wks (%) 80 % 93 % 98 % 96 %

Based on 13 subjects for the left hand pair and 16 subjects for the right hand pair of comparisons.

Table 3 Estimated folate intake of individual subjects against serum folate

Pre-supplementation folate intake Initial serum folate

Subject no. FFQ Diary ng=ml

15 601 415 5.25 16 627 375 7.55 5 441 313 3.50 10 354 391 7.45 18 328 328 8.40 3 417 226 6.20 4 391 226 4.90 17 431 183 5.90 14 307 238 4.85 2 295 219 8.45 1 300 206 6.05 12 241 217 3.65 19 212 202 5.05 6 206 162 5.05 20 201 138 8.45 13 142 120 7.95

Subjects are arranged in descending order of sum of FFQ ‡ diary folate intake estimates.

side). Folic acid was attached to the ED in a way that does Results not interfere with spontaneous reassociation of the enzyme Compliance was estimated to range between 92 and 100% fragments to form active enzyme. However, there is also in experiment (1); and between 94 and 99% in experiment bovine folate binding protein (BMFBP) in the kit. (2) and in experiment (3) with the six subjects it appeared When it binds to the folic acid attached to the ED, it to be 100%. There were no side effects to blood taking or to prevents ED binding to EA to form active b-galactosidase. the folic acid solutions and no signi®cant intercurrent The procedure involves ®rst adding the BMFBP to the illnesses. subject's serum. BMFBP binds with the folate in the serum, then ED is added. Some of the BMFBP has bound to the serum folate so there is less BMFBP available to bind to the folic acid attached to the ED; then EA is added. The ED not bound to the BMFBP is able to combine with EA for form Serum folate after three weeks compared with two weeks of active b-galactosidase. Activity of this b-galactosidase was folic acid supplementation measured by of o-nitrophenol-b-D-galactopyra- In experiments (1) and (2) serum folates were measured noside (ONGP) and read at 405 nm in a COBAS FARA after two weeks and then three weeks of taking each of the spectrophotometer. Serum folate was proportional to four doses of folic acid supplements. Mean serum folates ONGP hydrolysis. are shown in Table 2. With each supplement serum folates Each procedure was made in duplicate and all samples were higher at three weeks than at two weeks but the from an individual subject in an experiment were measured differences were only important at 100 and 500 mg supple- in the one run. The reliability and standardization of our mentation and only signi®cant at 100 mg folic acid. method was further checked by parallel determination of folates in the 18 sera from experiment (3) at the Haematol- ogy Laboratory, Westmead Hospital, using a completely different Access competitive-binding receptor assay (Sanof1 Diagnostics Pasteur Inc, USA). Agreement Correlation of folate intake estimations with serum folate between the two methods was consistently very close, before supplementation with our values on average 3% lower than those at In experiment (2) we estimated folate intake, both from a Westmead Hospital. In experiment (3), bloods were also food frequency questionnaire and from a 4 d diary. These taken into EDTA tubes and analysed for red cell folate by estimates and the starting serum folates of the 16 subjects Access competitive binding assay at the Haematology are shown in Table 3. Department, Westmead Hospital. There was some correlation between the two estimates The subjects were interested volunteers. Ethical of dietary folate intake (r2 ˆ 0.59) but, as can be seen in approval for the experiments was given by the University Table 3 there was no discernible correlation between folate of Sydney's Human Experimentation Ethics Committee. intake and serum folate before supplementation. Quantitative reponses of serum folate AS Truswell and S Kounnavong 842 Table 4 Individual serum folate concentrations, ng=ml, in all subjects who participated in any of the three experiments. To convert to nmol=l multiple by 2.23

Days 0 21 42 70 91 112 140 161 182

Serum folate on Serum folate on Serum folate on

Subject Initial 100 1000 Initial 500 1500 Initial 1000 2000 no. serum 1 mg=day mg=day serum 2 mg=day mg=day serum 3 mg=day mg=day

1 5.4 9 13.8 6.06 9.65 13.75 9.2 13 17.35 2 9.55 15.15 19.9 8.45 13.3 18.25 12.8 15.5 20.26 4 3.95 7.35 11.2 4.9 11.2 14.8 12.15 13.85 17.65 6 3.2 6.15 9.8 5.05 8.65 14.65 12.1 14.65 17.3 10 5.75 5.75 12.9 7.45 12.3 13.6 8.85 12.15 18.6 13 5.5 9.6 13.85 7.95 11.5 12.85 11.15 13.85 19.85 Mean 5.55 8.83 13.57 6.64 11.1 14.65 11.04 13.83 18.5 3 5.7 9.85 17.2 6.2 13.1 14.8 Ð Ð Ð 5 4.2 6.3 10.3 3.5 10.7 15.35 Ð Ð Ð 12 2.25 4.6 8.45 3.65 8.3 10 Ð Ð Ð 8 5.75 12.15 15.6 Ð Ð Ð Ð Ð Ð 9 5.45 6.85 13.5 Ð Ð Ð Ð Ð Ð 11 9.5 8.7 14.8 Ð Ð Ð Ð Ð Ð 7 1.6 5.45 12.6 Ð Ð Ð Ð Ð Ð 14 Ð Ð Ð 4.85 11 13.5 Ð Ð Ð 15 Ð Ð Ð 5.25 8.65 13.73 Ð Ð Ð 16 Ð Ð Ð 7.55 12.6 13.85 Ð Ð Ð 17 Ð Ð Ð 5.9 12.1 12.2 Ð Ð Ð 18 Ð Ð Ð 8.4 10.3 13.6 Ð Ð Ð 19 Ð Ð Ð 5.95 10.2 12.8 Ð Ð Ð 20 Ð Ð Ð 8.45 12.9 13.4 Ð Ð Ð Mean 5.18 8.22 13.38 6.17 11.03 13.84 11.04 13.83 18.50 all 20 (2.29) (0.83) (0.86) (0.72) (0.4) (0.4) (0.67) (0.48) (0.53) (s.e)

The left hand three columns are the data from experiment (1); the middle three columns are from experiment (2), and the right hand three columns are from experiment (3).

Table 5 The increases in serum folate, ng=ml, in six subjects, for each dose of folic acid supplementation (and expressed per 100 mg folic acid supplement)

Dose of folic acid (mg=d)

100 1000 500 1500 1000 2000 Subject no. d21±d0 d42±d0( Ä 10) d91±d0( Ä 5) d112±d0( Ä 15) d161±d0( Ä 10) d182±d0( Ä 20)

1 3.6 8.4 (0.84) 4.25 (0.85) 8.35 (0.55) 7.60 (0.76) 11.95 (0.59) 2 5.6 10.35 (1.03) 3.75 (0.75) 8.7 (0.58) 5.95 (0.59) 10.71 (0.53) 4 3.4 7.25 (0.75) 7.25 (1.45) 10.85 (0.72) 9.9 (0.99) 13.7 (0.68) 6 2.95 6.6 (0.66) 5.45 (1.09) 11.45 (0.76) 11.45 (1.41) 14.1 (0.7) 10 0 7.15 (0.71) 6.55 (1.34) 7.85 (0.52) 6.4 (0.64) 12.85 (0.64) 13 4.1 8.35 (0.83) 6.00 (1.20) 7.35 (0.49) 8.35 (0.83) 14.35 (0.71) Mean 3.28 8.01 (0.80) 5.54 (1.10) 9.09 (0.6) 8.27 (0.82) 12.94 (0.65)

Serum folates at different doses of folic acid supplementation Table 5 shows the increases in serum folate above the The serum folate results of all the individual subjects in original pre-supplementation (day 0) value for the six experiments (1), (2) and (3) are combined in Table 4. The subjects in experiments (1), (2) and (3). When the increase serum folates after folic acid supplements are the three of serum folate is expressed per 100 mg added folic acid, week values. The two week values are omitted because the the increment was maximal for the ®rst 100 mg then about new higher level of serum folate is not always reached by one-third of this to 1000 mg folic acid and about one-®fth two weeks (Table 2). between 1000 and 2000 mg folic acid supplement per day. The table has two rows of mean values for serum folate. The increases of serum folate, with each step-up of folic The upper row (after the top six subjects) is for the six acid dose were all signi®cant (at P < 0.05) by Fisher PLSD subjects who participated in all three experiments and took and mostly signi®cant by the more conservative Scheffe F- ®ve doses of folic acid (and 1000 mg taken in two courses). test. The bottom line gives the mean serum folates for all 20 women who participated in any of the three experiments. Effect of initial serum folate level on rise of serum folates Figure 1 shows the plot of mean serum folates for the six with folic acid supplementation subjects who took six three-week supplements of ®ve Our subjects did not start with serum folates all about the different doses of folic acid. The value for 1000 mg folic same value. They ranged from 1.60±9.55 ng=ml (nearly acid is the mean for the two periods. six-fold) in experiment (1) and from 3.50±8.45 ng=ml in Figure 2 shows the plot of mean serum folates for nine experiment (2). Figure 3 shows the high and low serum subjects who took four doses of folic acid for three weeks: folate subgroups (halves) of the subjects in experiment (2). 100, 500, 1000 and 1500 mg=d. Those who started with lower serum folates reached lower Quantitative responses of serum folate AS Truswell and S Kounnavonget al 843 Table 6 Folate concentrations in ng=ml (to convert to nmol=l multiply by 2.23)

Subject Pre-supplementation ‡ 1000 mg folic acid=d ‡ 2000 mg folic acid=d

Serum Red cell Serum Red cell Serum Red cell

1 9.2 341 13.0 377 17.35 406 2 12.8 373 15.5 394 20.3 539 4 12.15 376 13.85 386 17.65 539 6 12.1 315 14.65 392 17.3 404 10 8.85 451 12.15 447 18.6 456 13 11.15 490 13.85 469 19.85 449 mean 11.04 391 13.83 410 18.50 453 mean changes from start (%) ‡ 25% ‡ 5% ‡ 68% ‡ 16%

Figure 1 Mean serum folate ( Æ s.e.m.) for the six subjects who took six three weeks supplements of ®ve different doses of pure folic acid Figure 2 Mean serum folate ( Æ s.e.m.) for the nine subjects who took (Experiments (1), (2) and (3). four doses of pure folic acid in Experiments (1) and (2). Serum folate on usual diet is mean of both Experiments. serum levels with folic acid supplementation. The same phenomenon was seen in experiment (1).

Response of red cell folates In experiment (3) the effect on serum folate and on red cell folate of 1000 and 2000 mg folic acid supplements were compared. These are shown in Table 6. Red cell folate was not as sensitive or reliable in responding to three weeks of 1000 or 2000 mg of folic acid.

Discussion In our experiments, while individual subjects were not always consistent, the mean serum folate of small groups of women showed signi®cant increases in response to 100 mg supplements of pure folic acid. Clear increases of serum folate were found for every increase of folic acid supplement dosage, through 100, 500, 1000, 1500 and 2000 mg d. These increases were statistically signi®cant. It took three weeks for serum folate to reach the new Figure 3 Mean serum folate over supplementation periods, analyzed in higher level after increases in folic acid intake. Serum groups of low (  5.45 ng=ml) and high (  5.5 ng=ml) initial serum folate folates appeared to take longer to reach the new level levels. Data are Mean Æ SEM (for low serum folate n ˆ 7 and high serum after small increments in folic acid supplementation than folate n ˆ 6). after 1000 mg=d or more. Our ®ndings show that serum folate does not respond as quickly to increased intake as, would think that `recent' suggests in nutritional parlance for example, urinary nitrogen or sodium. In standard texts (geology has a different perspective). serum folate is said to re¯ect `recent dietary intake' but Three weeks was also the time for serum folate to fall three weeks for full response is longer than some people from an ordinary level to a diagnostically low level on a Quantitative reponses of serum folate AS Truswell and S Kounnavong 844 It may be helpful to compare our provisional plot for folic acid with the curve for recently reported by Levine et al (1996), based on strict control of the diet of seven men admitted to a clinical research unit for four to six months. Their plasma ascorbates were ®rst brought to the same low level by having the subjects on a depletion diet. The curve of plasma ascorbate rose steeply up to an intake of 100 mg ascorbic acid=d; it was curved through 200±400 mg ascorbic acid=d, and above this intake the plasma ascorbate plot was almost horizontal, only rising 9% from 1000 mg intake to 2500 mg ascorbic acid=d. Plasma ascorbate therefore rises very little with increasing ascorbic acid intakes in the supplement range. By contrast serum folate continues to rise steadilyÐit does not pla- teauÐinto the supplement intake range. Evidently the limited absorption and urinary overspill that prevent con- Figure 4 Mean serum folate over supplementation periods, analyzed in tinuing rise of plasma ascorbate, except in chronic renal groups of low (  5.90 ng=ml) and high (  6.05 ng=ml) initial serum folate levels. Data are Mean Æ SEM (for low serum folate n ˆ 8 and for failure (Allman et al, 1989) do not operate in folate high serum folate n ˆ 8). metabolism. Although we have studied women, there is no obvious reason why serum folates would behave differently in men, very low folate intake in Herbert's (1962) classic depletion though the numbers may be a little different. Tsui & experiment. We have now shown that it takes about the Nordstrom (1990) reported similar increases of serum same three weeks for serum folate to respond to a moderate folate in male and female adolescents given supplements increase of intake. Von der Porten et al (1992) reported a of 400 mg folic acid=d (mean serum folate 8.1 ng=ml on similar response time. They gave six healthy men 1600 mg usual diet to 12.9 ng=ml with 400 mg folic acid supple- deuterium-labelled folic acid for four weeks, serum folate ment). rose and reached a plateau after 18 d. We found no discernible correlation between The time for serum folate to settle back to its unsupple- estimated dietary folate and serum folate. Subjects' mented level after stopping high doses, however, appears to questionnaire replies and diary records were not interro- be well over four weeks. On withdrawing folic acid gated by a dietitian; there are no values for folate supplements subjects' serum folates had still not come in Australian foods, and folate's bioavailability in different back to their original levels after four weeks back on foods differ considerably and cannot be reliably usual diet with no supplementation. Mean serum folate of predicted (Gregory, 1995). We can be more con®dent the six repeating subjects was 5.5 ng=ml at the start (Table about the serum folate measurements. A set of them 4) but four weeks after stopping 100, then 1000 mg it was were validated by an independent laboratory using a still 6.64 ng=ml and then four weeks after stopping 500, different method. Six of our subjects took 1000 mg folic then 1500 mg folic acid=d it remained high at 11.04 ng=ml. acid for two periods of three weeks, and their mean For this reason the ®rst starting value (day 0) only was used serum folates differed by only 3% (8.01 and 8.27 mg=ml). in calculating increases of serum folate for each dose of It was impressive that in experiments (1) and (2) the supplement (in Table 5). There is a difference between the subgroup of subjects who started with lower serum folates time for serum folate to fall on changing from ordinary before supplementation showed similar rises with supple- dietary intakes to depletion (three weeks) and on changing mentation to the higher serum folate subgroup's but their down from supra-nutritional dosage (over 1000 mg=d) of serum folates tracked consistently lower (Figures 3 and 4). folic acid. In our experiments we have used pure folic acid (pteroyl The shape of the plot relating folic acid intake and serum ) but it could be objected that most dietary folate was steep for the ®rst 100 mg and thereafter was a less folate comes from other compounds, for example 5-methyl steep, almost linear upward slope that appeared to ¯atten tetrahydrofolate. The form we used, however, serves as a slightly above 1000 mg. We did not give supplements above well-de®ned standard of bioavailability and is the form in 2000 mg but there are reports of supra-normal (fasting) serum which people's folate intake is increased by forti®cation of folates with supplements of larger dose. In the MRC trial foods and taking supplements. (MRC Vitamin Study Research Group, 1991) the supplement Up to now serum folate has been used predominantly for was 4000 mg folic acid=d and the median serum folate of 538 clinical work and for diagnosing inadequate dietary folate. women (before they became pregnant) was 44 ng=ml (10th Our experiments show that serum folate can also be used for to 90th centiles 23±194 ng=ml). 44 ng=ml is roughly double public health research, to assess whether representative the value of 18.5 ng=ml which was our average for 2000 mg groups of the population are receiving optimal folate folic acid. Mooij et al (1993) reported surprisingly much intakesÐoptimal in the light of recent evidence that folate higher serum folates after 5000 mg folic acid=d in pregnant intakes above the RDA or RNI reduce the risk of foetal women after nine weeks' of supplementation (115 ng=ml). malformations and homocysteinaemia. This method can be There can be considerable discrepancy in serum folates used as well to monitor the population's response to folate between different methods and kits (Raiten & Fisher, forti®cation of foods, increased use of supplements and 1994). Whatever the reason for the very high serum folates advice to choose more food sources of folate. For this in this last report it appears that the plot of serum folate public health application we believe that serum folate will against folic acid intake would continue at much the same be less labour-intensive, more objective and reproducible gradient between 2000 and 4000 as we observed between and more biologically relevant than estimation of dietary 1000 and 2000 mg=d. folate. Indeed in countries without available data on food Quantitative responses of serum folate AS Truswell and S Kounnavonget al 845 folates serum folates can provide real data on samples of the Health and Welfare Canada (1993): Issues: Folic Acid, the Vitamin that population while estimates of their dietary folate have to use helps Protect against Neural Tube (Birth) Defects. Ottawa: Health Protection Branch. borrowed data that may not be real. Herbert V (1962): Experimental nutritional folate de®ciency in man. Trans. Assoc. Am. Physicians.75, 307±320. AcknowledgementsÐWe are grateful to Henneke Elzerman (Wageningen Holland B, Welch AA, Unwin ID, Buss DA, Paul AA & Southgate DAT University) and Elizabeth Leece for pilot experiments, to Dr N. Choudhury (1991): McCance and Widdowson's The Composition of Foods, 5th ed. for help with blood-taking, to Messrs. Z. Ahmad, Z. Khan, Bill Lowe Cambridge: Royal Society of Chemistry. and Dr M. Naghii for advice and help in the laboratory of the Human Kohlmeier L (1995): Future of dietary exposure assessment. Am. J. Clin. Nutr. 61 (suppl): 702S±709S. Nutrition Unit, to Dr Mark Hertzberg and Ms Helen Leung of Westmead Leece EA, Truswell AS, Elzerman H, Choudhury N & Ahmad ZI (1996): Hospital for red cell folates, and to Isa Hopwood who typed the manu- Relation between serum folate and folate intake in women. Proc. Nutr. script. Soc. Australia 19, 126. Levine M, Conry-Cantilena C, Wang Y, Welch RW, Washko PW, Dhariwal KR, Park JB, Lazarev A, Graumlich JF, King J, Cantilena LR (1996): Vitamin C in healthy volunteers: evidence References for a recommended dietary allowance. Proc. Natl. Acad. Sci. USA. 93, Allman MA, Truswell AS, Stewart PM, Yau DF, Tiller DJ, Horvath JS, 3704±3709. Duggan GG, Hall BM (1989): Vitamin supplementation of patients Medical Research Council Vitamin Study Research Group (1991): Pre- receiving haemodialysis. Med. J. Austral. 150, 130±133. vention of neural tube defects: results of the MRC Vitamin Study. Anonymous (1995): Acid folique et grossesse. Societe francaise de Lancet 338, 131±137. pediatrie Comite de nutrition. Arch. de Pediatrie. 2, 173±181. Mooij PNM, Steegers-Thennissen RPM, Thomas CMG, Doesburg MH & Bouskey CJ, Beresford SAA, Omenn GS & Motulsky AG (1995): A Eskes TKAB (1993): Periconceptual vitamin pro®les are not suitable quantitative assessment of plasma as a risk factor for for identifying women at risk for neural tube defects. J. . 123, 197± vascular disease. Probable bene®ts of increasing folic acid intakes. 203. JAMA 274, 1049±1057. Morrison HI, Schaubel D, Desmenles M & Wigle DT (1996): Serum folate Bower C (1996): Folate and the prevention of birth defects. Austral. J. and risk of fatal coronary heart disease. JAMA 275, 1893±1896. Nutr. Diet.53, S5±S8. National Health and Medical Research Council (1995): Folate Forti®ca- Burlingame BA, Milligan GC, Apimerica DE & Arthur JM (1993): The tion. Report of the Expert Panel on Folate Forti®cation. Canberra: Concise New Zealand Food Composition Tables. New Zealand Institute Australian Government Publishing Service. for Crop and Food Research, Palmerston North. Phillips DR & Wright AJA (1982): Studies on the response of Lactoba- Centers for Disease Control (1992): Recommendations for the use of folic cillus casei to different folate monoglutamates. Br. J. Nutr. 47, 183± acid to reduce the number of spina bi®da and other neural tube defects. 189. MMWR Morb. Mortal. Wkly. Rep. 41, 1±7. Raiten DJ & Fisher KD (eds) (1994): Assessment of Folate Methodology Cravo ML, Fidalgo P, Pereira AD, Gouveia-Oliveira A, Selhub CP & Used in the Third National Health and Nutrition Examination Survey Mason JB (1994): DNA as an intermediate biomarker in (NHANES III, 1988±1994). Bethesda MD, USA: Life Sciences colorectal cancer: modulation by folic acid supplementation. Eur. J. Research Of®ce, Federation of American Societies for Experimental Cancer Prev.3, 473±479. Biology. Czeizel A & Dudas I (1992): Prevention of ®rst occurrence of neural tube Sauberlich HE, Dowdy RP & Skala JH (1974): Laboratory Tests for the defects by periconceptual vitamin supplementation. New Eng. J. Assessment of Nutritional Status. Cleveland, Ohio: CRC Press. Med.327, 1832±1835. Scheelings P (1996): Prospects for measuring folates in Australian foods. Expert Advisory Group (1992): Folic Acid and the Prevention of Neural Austral. J. Nutr. Diet 53 S23±S28. Tube Defects. London, UK: Department of Health. Staatstoezicht op de Volksgesondheid (1993): Preventie neuraalbuis Food and Drug Administration, USA (1996): Press release, 29 February. defecten. Rijswijk, 19 November. Gibson RS (1990): Principles of Nutritional Assessment. New York and Truswell AS (1996): Prospects for folic acid forti®cation and its monitor- Oxford: Oxford University Press, pp 464. ing. Austral. J. Nutr. Diet 53, S17±S23. Giovanucci E, Stampfer MJ, Colditz GA, Rimm EB, Trichopoulos D & Tsui JC & Nordstrom JW (1990): Folate status of adolescents: effect of Roser BA (1993): Folate, and alcohol intake and risk of folic acid supplementation. J. Am. Diet Assn. 90, 1551±1556. colorectal adenoma. J. Natl. Cancer Inst. 85, 875±883. Von der Porten AE, Gregory JS, Toth JP, Cerda JJ, Curry SK & Bailey LB Gregory JF (1995). The bioavailability of folate. In Folate in Health and (1992): In vivo folate kinetics during chronic supplementation of Disease, ed LB Bailey, pp 195±235. New York and Basel: Marcel human subjects with deuterium-labelled folic acid. J Nutr. 122, 1293± Dekker. 1299.