This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation

Differential Relations Between Cognition and ¹⁵N Isotopic Content of Hair in Elderly People With Dementia and Controls

^a OPTIMA, Radcliffe Infirmary, Oxford, United Kingdom
^b Research Laboratory for Archaeology, University of Oxford, United Kingdom

Jonathan H. Williams, OPTIMA, Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, United Kingdom E-mail: jonathan.williams{at}pharmacology.ox.ac.uk.

	Abstract

Top Abstract Methods Results Discussion Appendix References

 
Background. Previous researchers have suggested that a vegetarian diet or one rich in fish may protect against Alzheimer's disease (AD). However, assessing diet is difficult in AD patients. ¹⁵N:¹⁴N isotopic ratios (¹⁵N) of body proteins can estimate long-term dietary habits in a way that does not depend on memory. ¹⁵N is high in people who eat a lot of fish and low in vegetarians.

Methods. To choose between the vegetarian and fish hypotheses of AD, we compared dietary questionnaire reports and ¹⁵N of hair samples from AD patients and controls.

Results. Patients' cognitive scores related directly to reported frequency of eating fish and to hair ¹⁵N_AIR, but inversely to reported frequency of eating beans. Homocysteine levels related inversely to hair ¹⁵N_AIR in controls, but not in patients. Dietary questionnaire reports accounted for slightly more variance in ¹⁵N_AIR in patients than controls. Therefore, our questionnaire assessed dietary habits as reliably for individuals with AD as for cognitively unimpaired controls.

Conclusions. A diet rich in fish may ameliorate AD, possibly by lowering homocysteine, but more vegetarian diets do not. In fact, eating beans correlated with worse cognition in AD patients. Further studies should test if restricting the intake of beans slows the progression of AD.

MANY researchers have proposed that dietary factors contribute to Alzheimer's disease (AD). On this view, diet and/or the absorption of nutrients should differ between patients with AD and controls. Most studies have assessed diet in AD by questionnaires ⁽¹⁾⁽²⁾. However, memory problems may limit such assessments. Our first goal was to assess the accuracy of dietary questionnaire reports in AD. To do this, we supplemented a dietary questionnaire with analyses of ¹⁵N:¹⁴N isotopic ratios (¹⁵N) in protein from elderly patients with AD and healthy controls.

¹⁵N provides a broad indication of dietary protein sources in a way that does not depend on memory. The technique is a common method of dietary assessment in ecology and archeology ^{(3)(4)(5)(6)(7)(8)}. ¹⁵N reflects the position of a dietary protein source in the food chain and can also distinguish between marine and terrestrial protein sources. It is low in vegetable and plant proteins, intermediate in common terrestrial animal proteins (milk, meat from ungulates), and higher in seafoods (fish, shellfish), because marine food chains are generally longer ⁽⁹⁾. Humans derive all their nitrogen from their diet, so a person's ¹⁵N integrates the ¹⁵N in their dietary sources ⁽¹⁰⁾ over long periods ⁽¹¹⁾. Consequently, people who habitually eat more fish have higher ¹⁵N, and vegans have lower ¹⁵N than those whose dietary protein is derived mainly from terrestrial animal (meat, dairy) foods ^{(11)(12)(13)(14)}. ¹⁵N is a nonspecific measure: quite different categories of foods may have similar ¹⁵N, and people who eat a balanced range of vegetable, dairy, and fish proteins should have the same ¹⁵N as those who eat only terrestrial meats. Nevertheless, ¹⁵N measures can help to assess if dietary questionnaire reports are as accurate for AD patients as for healthy aged controls.

Dietary theories of AD include proposals that AD patients are deficient in omega-3 fatty acids ⁽¹⁵⁾ or that a vegetarian diet protects against AD ⁽¹⁶⁾. ¹⁵N indexes an individual's position on the food chain (see above). So, the hypothesis that a vegetarian diet protects against AD ⁽¹⁶⁾ predicts that patients with AD should have higher ¹⁵N. Conversely, the hypothesis ⁽¹⁵⁾ that fish oils protect against AD predicts that AD patients would have lower ¹⁵N. Our second goal was to measure ¹⁵N to choose between these competing hypotheses.

High homocysteine is a risk factor for AD ^(17)(18). Homocysteine levels depend partly on diet. Foods containing methionine increase homocysteine ^{(19)(20)(21)(22)}, but those containing folate (e.g., fresh fruit) ^(23)(24) or vitamin B₁₂ (e.g., meat) ⁽¹⁹⁾ lower it. This raises the question whether high homocysteine levels in patients with AD reflect differences in their diet or in their metabolism of folate and methionine. Our third goal was to choose between these alternatives, by testing whether homocysteine showed similar relations to ¹⁵N in AD patients and controls.

Our final goal was to test if dietary factors relate to the severity of AD. If dietary factors modulate the progression of AD, then establishing such relations would be an important step toward interventional studies of dietary constituents in AD.

	Methods

Top Abstract Methods Results Discussion Appendix References

 
The Oxford Project To Investigate Memory and Ageing (OPTIMA) is a longitudinal study of people with dementia and age-matched controls. Jobst and his colleagues ⁽²⁵⁾ have described the OPTIMA protocol previously. In brief, participants underwent neuropsychological assessment (CAMCOG: see ⁽²⁶⁾) and had a medical examination, blood tests, and radiological scans to see if they met criteria for AD (NINCDS-ADRDA; ⁽²⁷⁾). Additionally, they provided hair samples (cut adjacent to the scalp) for mass spectrometric analysis of ¹⁵N (see below).

Participants (or carers) also completed a semiquantitative dietary questionnaire (see Appendix) to assess their average frequency of eating meat, fish, dairy products, maize, staples (bread, pasta, or rice), and fresh fruit and vegetables. Our questionnaire asked for the frequency of eating different foods in a manner similar to that validated by a large trial ^(28)(29).

Nitrogen isotopic analysis of hair samples was as detailed in O'Connell and Hedges ⁽¹¹⁾. Samples were cleaned using organic solvents and water to remove dirt, lipid, or shampoo residues, then wrapped lengthways in aluminum foil, cut into 1.5-cm sections, dried under vacuum, and rolled into balls. Isotopic analyses were performed using an automated elemental analyzer (Carlo Erba, Milan, Italy) coupled to a Geo 20/20 isotope ratio mass spectrometer (PDZ Europa, Crewe, United Kingdom). We corrected results for mass dependence by comparison with laboratory standards run in conjunction with the samples. All results in this article are reported using the ‘’ notation in units of permil () relative to the international standard AIR ⁽³⁰⁾: ¹⁵N_AIR = [(^15/14N_sample /^15/14N_AIR) - 1] x 1000. We analyzed each sample in duplicate: replicate measurement errors on standards were less than ±0.2%. Variations in hair pigmentation and treatment processes do not affect isotopic analyses ^(11)(14).

We analyzed differences in single variables between patient and control groups using nonparametric statistics: ², Wilcoxon-Mann-Whitney z (WMW z). The dietary questionnaire reports were only semiquantitative, so we analyzed their relations to other variables using polychoric correlations (which give unbiased estimates of relationship for variables of this kind) ⁽³¹⁾.

Homocysteine levels depend on age, gender, and levels of folate and vitamin B₁₂ ⁽³²⁾. The mean red cell volume (MCV) provides a biological marker of vitamin B₁₂ and homocysteine levels ⁽³³⁾. We tested the dependence of homocysteine upon ¹⁵N_AIR and the other variables using multiple linear regression. The regression first fitted the main effect of each independent variable. It then fitted the interactions of these variables with patient status. Finally, it dropped any nonsignificant interactions or main effects. We analyzed the dependence of cognitive performance (CAMCOG scores) upon the same variables using a similar multiple regression that included homocysteine as an independent variable.

	Results

Top Abstract Methods Results Discussion Appendix References

 
Sample Characteristics
We obtained hair samples from 110 community-dwelling older people (mean age was 74.1 ± 8.3 years, range 57–92; 49 men). Forty-three had clinical diagnoses of AD (28 NINCDS "Possible" and 15 NINCDS "Probable"; median CAMCOG score 74, interquartile range 50–85), and 67 were controls who had no objective memory problems (all NINCDS "Negative"; median CAMCOG score 100, interquartile range 96–103). The age and gender distributions of patient and control groups did not differ (gender: ² = 1.8, 1 df, p > .1; age: WMW z = 1.35, p = .18). Overall, the mean ¹⁵N_AIR was 9.02% ± 0.40 (SD), and the mean homocysteine level was 12.84 ± 4.87 (SD) µmol/l.

Relations Between ¹⁵N_AIR and Dietary Questionnaire Reports in Patients and Controls
Overall, ¹⁵N_AIR correlated positively with reported frequency of eating fish (polychoric correlation = 0.39, n = 107, p = .001) (Fig. 1), but negatively with that of eating beans (polychoric correlation = -0.29, n = 107, p = .02) (Fig. 2). Analysis of variance with polynomial trends showed that ¹⁵N_AIR fell with more frequent reported bean consumption in controls (F = 7.80; 1/101 df; p = .006), but fell even more sharply in patients (Patient x Beans interaction: F = 4.13; 1/101 df; p = .044) (see Fig. 2).

View larger version (20K): [in this window] [in a new window]   Figure 1. The dependence of 15NAIR (d15N - ordinate) upon the reported frequency of eating fish (abscissa) in the controls () and Alzheimer's disease (AD) patients (*). Frequency of eating defined as 0 = never; 1 = rarely; 2 = once a week; 3 = two or three times per week; 4 = every day. As expected, 15NAIR related directly to fish consumption in both groups.

View larger version (21K): [in this window] [in a new window]   Figure 2. The dependence of 15NAIR (d15N - ordinate) upon the reported frequency of eating beans (abscissa) in the controls () and Alzheimer's disease (AD) patients (*). Frequency of eating defined as 0 = never; 1 = rarely; 2 = once a week; 3 = two or three times per week; 4 = every day. As expected, 15NAIR related inversely to frequency of eating beans in both groups. Unexpectedly, the inverse relation was steeper in AD patients than controls.

Relations of AD to Dietary Questionnaire Preferences and ¹⁵N_AIR

Relation of Homocysteine to ¹⁵N_AIR
Patients' homocysteine levels were higher than controls' (WMW z = 2.60, p = .009) (as reported previously in the OPTIMA cohort [17]). This remained significant (t = 1.97, p = .05) in the multiple regression that covaried serum folate (t = 5.75, p < .001), age (t = 2.93, p = .004), and mean cell volume (t = 2.64, p = .01). In this final multiple regression model, two interactions were significant. First, homocysteine levels related inversely to ¹⁵N_AIR (t = -2.39, p = .019) in controls, but not in the patients (¹⁵N_AIR x Patient interaction: t = 2.02, p = .046) (Fig. 3). Second, homocysteine related directly to MCV in the controls (t = 3.54, p = .001), but not in the patients (MCV x Patient interaction: t = 2.30; p = .024) (not shown).

View larger version (23K): [in this window] [in a new window]   Figure 3. The relations between homocysteine (ordinate) and 15NAIR (d15N - abscissa) in the controls () and Alzheimer's disease (AD) patients (*). Homocysteine levels related inversely to 15NAIR in controls, but not in patients.

Dependence of Severity of AD on ¹⁵N_AIR and Dietary Questionnaire Reports

View larger version (23K): [in this window] [in a new window]   Figure 4. The relation between CAMCOG scores (ordinate) and 15NAIR in hair (abscissa) in the controls () and Alzheimer's disease (AD) patients (*). AD patients with higher 15NAIR (reflecting greater consumption of fish and/or lower consumption of beans) had higher CAMCOG scores.

View larger version (20K): [in this window] [in a new window]   Figure 5. The relation between CAMCOG scores (ordinate) and the reported frequency of eating beans (abscissa) in the controls () and Alzheimer's disease (AD) patients (*). Frequency of eating defined as 0 = never; 1 = rarely; 2 = once a week; 3 = two or three times per week; 4 = every day. AD patients who reported eating beans more frequently had lower CAMCOG scores.

	Discussion

Top Abstract Methods Results Discussion Appendix References

 
Our study provided novel tests of hypotheses that dietary variables contribute to AD. Patients and controls did not differ, overall, in their reported frequencies of consuming different foods. However, they did show different relations between those reports and laboratory indices. These findings suggest that patients absorb or metabolize important dietary compounds differently from controls. Finally, there were striking correlations between dietary reports and cognitive impairments in AD patients. Overall, our results were broadly consistent with the view that dietary factors contribute to AD.

¹⁵N_AIR was higher in people who reported eating fish more frequently (see Fig. 1), exactly in line with observations from previous studies ^(12)(13)(14). ¹⁵N_AIR was also lower in those who reported eating beans more frequently (Fig. 2). Beans and legumes derive from nitrogen-fixing plants, which have lower ¹⁵N_AIR than other plants ^(34)(35), so eating more beans should lower ¹⁵N_AIR, as we found. These results, therefore, indicate that our food frequency questionnaire tapped meaningful differences in individuals' dietary preferences. In fact, the dietary questionnaire responses accounted for 15–20% of the variance in ¹⁵N_AIR. At first, this may suggest that the dietary questionnaire reports were inaccurate. However, several factors could dilute the correlation between ¹⁵N_AIR and the questionnaire responses. First, our questionnaire asked about the frequency of eating different types of foods, which relates imperfectly to the quantities eaten ⁽²⁹⁾. Second, most preprocessed foods (e.g., biscuits) contain many ingredients, which participants probably did not take into account in their questionnaire responses. Third, ¹⁵N is not very food-specific (see introduction), and individual variations in ¹⁵N can only reflect relatively extreme dietary propensities. This limitation dilutes the relationship between the dietary reports and ¹⁵N and makes ¹⁵N unsuitable for testing if eating proteins from terrestrial animals relates to AD. Fourth, hair grows at about 1 cm per month ⁽³⁶⁾, but its ¹⁵N is an integrated average of the isotopic composition of an individual's diet over the majority of the year prior to its growth ^(11)(37). This is an advantage of our isotopic analyses over the dietary questionnaire responses: the latter are more likely to reflect participants' dietary intake over only the previous few days or weeks. In light of these limitations, the fact that the dietary questionnaire responses accounted for 15–20% of the variance in ¹⁵N_AIR is more impressive. We conclude that, overall, ¹⁵N_AIR and our questionnaire both reliably tap individuals' enduring dietary preferences. The two methods are complementary—the questionnaire emphasizes the frequency of eating different foods, but ¹⁵N probably relates more strongly to the quantities eaten.

Reliability of Dietary Questionnaires in AD
Our findings indicate that questionnaire reports of dietary habits from patients with AD are just as good as, or even better than, controls. Dietary questionnaire reports accounted for slightly more variance in ¹⁵N_AIR in patients (R² = .19) than controls (R² = .15). Probably, patients could not recall their diet, and their carers supplied much of this information. Such second-hand information should be less reliable than the first-hand information from controls. However, it is hard to see how such information dilution could explain how ¹⁵N_AIR related (inversely) to the frequency of eating beans more strongly in patients than controls (Fig. 2). Patients with AD may prefer more sweet foods ⁽³⁸⁾, but this cannot readily account for their stronger relation between ¹⁵N_AIR and their reported consumption of beans, nor could institutionalization, because all patients were community-dwelling. Therefore, we conclude that our questionnaire assessed dietary habits as reliably for individuals with AD as for cognitively unimpaired controls.

Implications of Our Results for Vegetarian and Dietary Fish Hypotheses of AD
Our second goal was to choose between the competing hypotheses that either a vegetarian diet ⁽¹⁶⁾ or a diet rich in fish ⁽¹⁵⁾ may protect against AD. Although correlational, our findings militate against the vegetarian hypothesis. Instead, they offered slight support to the fish hypothesis ⁽¹⁵⁾, because the AD patients took a slightly lower proportion of their animal protein as fish, compared to controls. The fact that the overall ¹⁵N distributions were similar for patients and controls is not consistent with either of the hypotheses that fish or vegetarian diets are protective against AD. However, after accounting for bean consumption, patients who reported eating less fish had lower CAMCOG scores. This was not merely due to better recall of eating fish in patients with lesser cognitive impairments, because patients' CAMCOG scores also related directly to their ¹⁵N (providing independent evidence that they ate more fish protein). Hence, possibly, a diet rich in fish may slow the progression of AD. However, our observational study cannot distinguish if reduced frequency of eating fish by more impaired patients is a cause or effect of their memory problems. Further experimental studies should test this.

Implications of Our Results for the Homocysteine Hypothesis of AD
Our findings suggest that high homocysteine levels in AD patients ^(17)(18) relate to changes in metabolism, rather than differences in diet. Homocysteine levels related to ¹⁵N_AIR only in controls (Fig. 3). This indicates that controls who ate more animal protein, especially fish, had lower homocysteine. This is consistent with previous observational studies ⁽¹⁹⁾ and with evidence that dietary supplements of fish oils can lower homocysteine ⁽³⁹⁾. In contrast to controls, homocysteine did not relate to ¹⁵N_AIR in patients. This suggests that patients' higher homocysteine resulted from changes in their absorption or metabolism of methionine, folate, or B₁₂, rather than their diet. Further studies should confirm our findings and test if dietary modifications or supplements can lower homocysteine levels in AD.

Relationship of Dietary Reports to Severity of AD
Our correlational results indicated that dietary factors relate to the severity of AD, but they cannot illuminate the mechanism of this relationship. AD is chronic, so that although ¹⁵N_AIR integrates information about dietary habits over the preceding year, the disease started before that time. We are thus unable to know if the dietary changes that we found are causes, effects, or correlates of the severity of AD. (For example, the main maize-based food that people eat in this country is breakfast cereal, which may be their main source of folate ⁽⁴⁰⁾. So, the correlation of eating maize-based foods with CAMCOG may reflect effects of folate on CAMCOG [17], rather than any quality intrinsic to maize.) However, our results suggest a need for interventional studies to test the possibility that dietary factors may modulate the severity of AD. In particular, patients who reported eating more beans had lower CAMCOG scores (Fig. 5). The fact that the controls showed a similar, though not quite significant, trend reinforces this unexpected result and tends to exclude the possibility that participants ate more beans due to impaired cognition. The finding that CAMCOG correlates with reported consumption of beans, but not other plant foods, is not easy to explain. One possibility is that beans contain a higher level of antinutrients, such as phytates, that could impair the absorption or metabolism of other micronutrients. Further studies should confirm our findings in population-derived samples and test if restricting the frequency of eating beans prevents cognitive deterioration in patients with AD.

	Acknowledgments

 
Dr. Williams received financial support from the Takayama Foundation, and Dr. O'Connell received financial support from the Wellcome Trust.

We thank the participants in OPTIMA who completed the dietary questionnaire and provided the hair samples. We also thank the OPTIMA nurses for obtaining the hair samples. We thank the editor and an anonymous reviewer for constructive criticism of the manuscript.

Received February 21, 2002

Accepted June 6, 2002

	Appendix

Top Abstract Methods Results Discussion Appendix References

 

	References

Top Abstract Methods Results Discussion Appendix References

1. Forster DP, Newens AJ, Kay DW, Edwardson JA, 1995. Risk factors in clinically diagnosed presenile dementia of the Alzheimer type: a case-control study in northern England. J Epidemiol Community Health 49:253-258. [Abstract/Free Full Text]
2. Renvall MJ, Spindler AA, Ramsdell JW, Paskvan M, 1989. Nutritional status of free-living Alzheimer's patients. Am J Med Sci 298:20-27. [Medline]
3. Best PB, Schell DM, 1996. Stable isotopes in southern right whale (Eubalaena australis) baleen as indicators of seasonal movements, feeding and growth. Marine Biology 124:483-494.
4. Schoeninger MJ, Iwaniec UT, Nash LT, 1998. Ecological attributes recorded in stable isotope ratios of arboreal prosimian hair. Oecologia 113:222-230.
5. Hobson KA, 1999. Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia 120:314-326.
6. Bonsall JC, Lennon RJ, McSweeny K, et al. 1997. Mesolithic and Early Neolithic in the Iron Gates: a palaeodietary perspective. J Eur Archaeol 5:50-92.
7. Richards MP, Hedges REM, Molleson TI, Vogel JC, 1998. Stable isotope analysis reveals variations in human diet at the Poundbury Camp cemetery site. J Archaeol Sci 25:1247-1252.
8. Katzenberg MA, Weber A, 1999. Stable isotope ecology and palaeodiet in the Lake Baikal region of Siberia. J Archaeol Sci 26:651-659.
9. Schoeninger MJ, DeNiro MJ, 1984. Nitrogen and carbon isotopic composition of bone collagen from marine and terrestrial animals. Geochim Cosmochim Acta 48:625-639.
10. DeNiro MJ, Epstein S, 1981. Influence of diet on the distribution of nitrogen isotopes in animals. Geochim Cosmochim Acta 45:341-351.
11. O'Connell TC, Hedges REM, 1999. Investigations into the effect of diet on modern human hair isotopic values. Am J Phys Anthropol 108:409-425. [Medline]
12. Schoeller DA, Minagawa M, Slater R, Kaplan IR, 1986. Stable isotopes of carbon, nitrogen and hydrogen in the contemporary North American human food web. Ecol Food Nutr 18:159-170.
13. Minagawa M, 1992. Reconstruction of human diet from ¹³C and ¹⁵N in contemporary Japanese hair: a stochastic method for estimating multi-source contribution by double isotopic tracers. Appl Geochem 7:145-158.
14. Yoshinaga J, Minagawa M, Suzuki T, et al. 1996. Stable carbon and nitrogen isotopic composition of diet and hair of Gidra-speaking Papuans [published erratum appears in Am J Phys Anthropol. 1996;100:611]. Am J Phys Anthropol 100:23-34. [Medline]
15. Horrocks LA, Yeo YK, 1999. Health benefits of docosahexaenoic acid (DHA). Pharmacol Res 40:211-225. [Medline]
16. Giem P, Beeson WL, Fraser GE, 1993. The incidence of dementia and intake of animal products: preliminary findings from the Adventist Health Study. Neuroepidemiology 12:28-36. [Medline]
17. Clarke R, Smith AD, Jobst KA, Refsum H, Sutton L, Ueland PM, 1998. Folate, vitamin B12, and serum total homocysteine levels in confirmed Alzheimer disease. Arch Neurol 55:1449-1455. [Abstract/Free Full Text]
18. McCaddon A, Davies G, Hudson P, Tandy S, Cattell H, 1998. Total serum homocysteine in senile dementia of Alzheimer type. Int J Geriatr Psychiatry 13:235-239. [Medline]
19. Mann NJ, Li D, Sinclair AJ, et al. 1999. The effect of diet on plasma homocysteine concentrations in healthy male subjects. Eur J Clin Nutr 53:895-899. [Medline]
20. Pieniazek D, Grabarek Z, Rakowska M, 1975. Quantitative determination of the content of available methionine and cysteine in food proteins. Nutr Metab 18:16-22. [Medline]
21. Reis RP, Azinheira J, Reis HP, Pina JE, Correia JM, Luis AS, 1999. The effect of sex and menopause on basal blood levels of homocysteine and after methionine loading. Rev Port Cardiol 18:155-159. [Medline]
22. Stolzenberg SRZ, Miller ER, III Maguire MG, Selhub J, Appel LJ, 1999. Association of dietary protein intake and coffee consumption with serum homocysteine concentrations in an older population. Am J Clin Nutr 69:467-475. [Abstract/Free Full Text]
23. Appel LJ, Miller ER, III Jee SH, et al. 2000. Effect of dietary patterns on serum homocysteine: results of a randomized, controlled feeding study. Circulation 102:852-857. [Abstract/Free Full Text]
24. Brouwer IA, van Dusseldorp M, Thomas CM, et al. 1999. Low-dose folic acid supplementation decreases plasma homocysteine concentrations: a randomized trial. Am J Clin Nutr 69:99-104. [Abstract/Free Full Text]
25. Jobst KA, Smith AD, Szatmari M, et al. 1992. Detection in life of confirmed Alzheimer's disease using a simple measurement of medial temporal lobe atrophy by computed tomography. Lancet 340:1179-1183. [Medline]
26. Roth M, Tym E, Mountjoy CQ, et al. 1986. CAMDEX. A standardised instrument for the diagnosis of mental disorder in the elderly, with special reference to the early detection of dementia. Br J Psychiatry 149:698-709. [Abstract/Free Full Text]
27. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM, 1984. Clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer's Disease. Neurology 34:939-944. [Abstract/Free Full Text]
28. Kroke A, Klipstein Grobusch K, Voss S, et al. 1999. Validation of a self-administered food-frequency questionnaire administered in the European Prospective Investigation into Cancer and Nutrition (EPIC) Study: comparison of energy, protein, and macronutrient intakes estimated with the doubly labeled water, urinary nitrogen, and repeated 24-h dietary recall methods. Am J Clin Nutr 70:439-447. [Abstract/Free Full Text]
29. Riboli E, Kaaks R, 1997. The EPIC Project: rationale and study design. European Prospective Investigation into Cancer and Nutrition. Int J Epidemiol 26:6-14.
30. Mariotti A, 1983. Atmospheric nitrogen is a reliable standard for natural ¹⁵N measurements. Nature 303:685-687.
31. Joreskog KG, Sorbom D. LISREL 8 User's Reference Guide. Chicago, IL: Scientific Software International; 1993.
32. Langman LJ, Cole DE, 1999. Homocysteine: cholesterol of the 90s?. Clin Chim Acta 286:63-80. [Medline]
33. Curtis D, Sparrow R, Brennan L, Van der Weyden MB, 1994. Elevated serum homocysteine as a predictor for vitamin B₁₂ or folate deficiency. Eur J Haematol 52:227-232. [Medline]
34. Virginia RA, Delwiche CC, 1982. Natural ¹⁵N abundance of presumed N₂-fixing and non N₂-fixing plants from selected ecosystems. Oecologia 54:317-325.
35. DeNiro MJ, Hastorf C, 1985. Alteration of ¹⁵N/¹⁴N and ¹³C/¹²C ratios of plant matter during the initial stages of diagenesis: studies utilizing archaeological specimens from Peru. Geochim Cosmochim Acta 49:97-115.
36. Saitoh M, Uzuka M, Sakamoto M, Kobori T, 1969. Rate of hair growth. Montagna W, Dobson R, , ed.Hair Growth 183-201. Pergamon Press, Oxford, England.
37. Jones RJ, Ludlow MM, Troughton JH, Blunt CG, 1981. Changes in natural carbon isotope ratio of the hair from steers fed diets of C₄, C₃ and C₄ species in sequence. Search 12:85-87.
38. Mungas D, Cooper JK, Weiler PG, Gietzen D, Franzi C, Bernick C, 1990. Dietary preference for sweet foods in patients with dementia. J Am Geriatr Soc 38:999-1007. [Medline]
39. Olszewski AJ, McCully KS, 1993. Fish oil decreases serum homocysteine in hyperlipemic men. Coron Artery Dis 4:53-60. [Medline]
40. Tucker K, Selhub J, Wilson P, Rosenberg I, 1996. Dietary intake pattern relates to plasma folate and homocysteine concentrations in the Framingham Heart Study. J Nutr. 126:3025-3031.

This article has been cited by other articles:

				E. Nurk, C. A Drevon, H. Refsum, K. Solvoll, S. E Vollset, O. Nygard, H. A Nygaard, K. Engedal, G. S Tell, and A D. Smith Cognitive performance among the elderly and dietary fish intake: the Hordaland Health Study Am. J. Clinical Nutrition, November 1, 2007; 86(5): 1470 - 1478. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation