Plasma Total-Homocysteine in Anorexia Nervosa

Plasma total-homocysteine in anorexia nervosa

D Moyano1, MA Vilaseca1. R Artuch1, C Valls1 and N Lambruschini2

1Serveis de BioquõÂmica i, 2Pediatria, Hospital Universitari Sant Joan de DeÂu, Barcelona, Spain

Objective: The measurement of plasma total-homocysteine (tHcy) as a marker of folate and cobalamin status in patients with anorexia nervosa. Design: Plasma tHcy, folate, cobalamin and other determinats of tHcy of a random group of patients with anorexia nervosa were compared with our own reference values. Setting: The study was performed at the tertiary children's Hospital Sant Joan de DeÂu. Sujects: All the female adolescents (n 43) coming to the Hospital during a one-year period, who were diagnosed with anorexia nervosa. Reference values for tHcy were simultaneously performed with apparently healthy adolescents (by history and analytical data) who underwent presurgical analysis for minor interventions, and other magnitudes we used our own reference values. Interventions: Plasma tHcy was measured by reverse fase HPLC with ¯uorescence detection of the SBDF derivatives. Folate and cobalamin concentrations were determined by radioimmunoassay. Results: tHcy was signi®cantly increased in anorexic patients compared to reference values (Mann±Whitney, P < 0.0001±0.001). Values were above reference range in 34% of patients, and high-normal range in 53% of patients. tHcy concentrations lowered in 8 and 11 patients after nutritional rehabilitation. Cobalamin and folate were in the reference range except for one case. No signi®cant correlation was found among tHcy, vitamins and other determinants of tHcy concentration. Conclusions: tHcy concentrations appear signi®cantly increased in adolescents with anorexia nervosa, most probably owing to subclinical folate de®ciency. This might be caused by both, intracellular folate de®ciency and impaired availability. Abnormal plasma tHcy values were completely corrected after nutritional rehabilitation. Sponsorship: D Moyano was the recipient of a grant from Hospital Sant Joan de DeÂu, Barcelona. Descriptors: Anorexia nervosa; folate; cobalamin; total-homocysteine; malnutrition

Introduction homocysteine (tHcy), the sum of all the homocysteine species that release this amino acid by reduction, increases Anorexia nervosa is a severe nutritional disorder, affecting considerably when there is an intracellular de®ciency of mostly female adolescents, characterized by excessive food folate or cobalamin (Ueland et al, 1993). restriction that results in a marked loss of weight. Although Our aim in this study was the detection of possible diets (self-designed and strictly controlled by the anorectic de®ciencies of folate and cobalamin in anorexia nervosa by patients) are not uniform, as is the case of macrobiotic the measurement of plasma tHcy. Cobalamin and folate nutrition (Schneede et al, 1994), the experience of several concentrations as well as other tHcy determinants (Refsum groups conclude that there is a general trend to a marasmic et al, 1997) were also determined in an attempt to ®nd a malnutrition in anorexia nervosa (Madruga et al, 1993; relationship with tHcy values. Nussbaum, 1992). Therefore, either protein and energy, as well as micronutrients may result defective in this condi- tion. However, in spite of being a severe eating disorder, it Materials and methods is associated with rather mild abnormalities in the classical parameters of malnutrition (Schebendach et al, 1992; Subjects Nussbaum, 1992). Studies of biochemical data on vitamin The female adolescents (n 43) coming to the Department and mineral status found in the literature are fragmentary of Pediatrics of our hospital during a one-year period, who and contradictory and most often reported for adults, were diagnosed of anorexia nervosa according to the however, subclinical vitamin de®ciencies might be consid- criteria of the American Psychiatric Association (1994). ered in this disorder (Fisher et al, 1995). The search for new They were analysed at diagnosis as a part of the biochem- diagnostic tools to detect the defective dietary components ical evaluation of their nutritional status. Body mass index of diet may help to apply the adequate treatment that avoids (BMI) calculated as weight divided by the square of height the neurologic sequelae involved in these vitamin de®cien- (kg=m2) was 17 (12±19) (median (range); normal popula- cies (Carmel, 1989). tion > 20). Renal function was normal in all the patients. Folate and cobalamin (as methylcobalamin) are Reference values for tHcy were simultaneously performed involved as substrate and coenzyme, respectively, in the with apparently healthy adolescents (by history and analy- remethylation of homocysteine to methionine. Plasma total tical data) who underwent presurgical analysis for minor interventions (n 90). For the other biochemical and hematological magnitudes we used our own age-matched reference values. Blood was collected again in a group of Correspondence: Dr MA Vilaseca, Servei de BioquõÂmica, Hospital Universitari Sant Joan de DeÂu, P. Sant Joan de DeÂu 2, 08950 Esplugues 11 patients after six months of nutritional rehabilitation (Barcelona), Spain. without special vitamin supplementation. Samples of Received 4 April 1997; revised 29 August 1997; accepted 5 October 1997 patients and controls used for reference values were Plasma total-homocysteine in anorexia nervosa D Moyano et al 173 obtained in accordance with the Helsinki Declaration of Table 1 Plasma tHcy, folate, cobalamin, methionine, cystine, MCV, 1975, as revised in 1983. hemoglobin, creatinine and BMI in anorexia nervosa Anorexia nervosa Reference values Specimens Fasting venous blood was collected from patients with Units Median (interval) N Median (interval) N anorexia nervosa and controls used for reference values at tHcy mmol=L 8:00 am. The samples were collected in Venoject tubes A (11±15 y) 9.5*** (5.9±24.1) 27 6.6 (4.4±10.8) 59 containing EDTA, placed in melting ice, and centrifuged B (16±18 y) 11.2** (6.3±30.1) 16 8.1 (4.6±11.3) 31 within 15 min (3000 g, 10 min at 4C). The plasma was Folate nmol=L 13.6 (3.2±45) 43 13.8 (5.6±39) 52 Cobalamin pmol=L 350 (186±931) 43 306 (175±809) 52 rapidly separated, frozen at 7 40 C and analysed within Cystine mmol=L 29** (14±84) 43 35 (11±83) 52 the next fortnight. Serum samples collected in parallel for Methionine mmol=L 30* (15±57) 43 23 (8±67) 52 nutritional control were used for cobalamin and folate MCV fL 90 (71±100) 43 88 (74±102) 52 measurements. Hemoglobin g=dL 13.2 (12±15) 43 13.5 (12±16) 52 Creatinine mmol=L 98*** (83±110) 43 67 (42±112) 52 BMI kg=m2 17 (12±19) 43 > 20 30 Methods Plasma samples were analysed for tHcy by reverse fase Mann-Whitney, ***P < 0.0001, **P < 0.001; *P < 0.02. HPLC with ¯uorescence detection of the SBDF derivatives (Vilaseca et al, 1997). Within-run imprecision (n 18): 3% (70.9 Æ 2.2 mmol=L), and 4% (8.9 Æ 0.3 mmol=L); between-day imprecision (n 18): 6% (6.9 Æ 0.4 mmol=L), and 6% (55.4 Æ 3.1 mmol=L). Cobalamin and folate were determined by radioimmunoassay (ICN Phar- maceuticals SimulTRAC, Costa Mesa, CA, USA). Folate within-run imprecision (n 20): 3.7% (5.5 Æ 0.4 nmol=L); between-day imprecision (n 20): 5.2% (5.1 Æ 0.5 nmol=L). Cobalamin within-run imprecision (n 20): 6.4% (318 Æ 40 pmol=L); between-day imprecision (n 20): 4% (320 Æ 54 pmol=L). Hematological indices: median corpuscular volum (MCV fL) and hemoblobin (g=dL), were measured with the autoanalyser Cobas Vega Roche (Diagnostic systems). Creatinine was measured by an automated adaptation (Olympus AU-510, Merck-Igoda S.A.) of JaffeÂ method (FarreÂ et al, 1991). As an additional study, plasma sulfur amino acids, methionine and cystine, were included in the results (measured by ion exchange chromatography with post-column ninhydrin detection) (Moyano et al, 1996).

Statistical analysis Kolmogorov±Smirnov test was applied to study the dis- tribution. Since tHcy, vitamin and amino acid values were Figure 1 tHcy values in anorexia nervosa (AN) distributed in two age not normally distributed, non-parametric Mann±Whitney groups (A) 11±15 y, and (B) 16±18 y, compared with age-matched test was used to compare patient and reference values, and reference values (RVa and b). Horizontal lines indicate the medians in all groups. Wilcoxon test to compare tHcy in the group of patients analysed at diagnosis and after nutritional rehabilitation. The Spearman linear regression coe®cient was used to established the relationship between concentrations of tHcy, folate, cobalamin, creatinine, BMI, methionine, cystine, MCV and hemoglobin in the patient group. Statis- tical analysis was performed with Statgraphics Statistical Graphics System, version 6.0.

Results Plasma tHcy increases with age, so patients were distributed in two aged groups: (A) 11±15 y, and (B) 16±18 y, according to our reference values. tHcy was signi®cantly higher in patients in comparison with the age-matched reference population (Mann±Whitney, P < 0.0001±0.001) (Table 1) (Figure 1). In seven patients of group A (26%) and eight of group B (50%) tHcy values were above the reference range. Moreover, 23 and of 43 patients (53%) showed tHcy concentration in the high-normal range of the reference values. tHcy concentrations lowered in 8 and of 11 patients after nutritional rehabilitation (Figure 2). Figure 2 tHcy concentrations at diagnosis (1) of anorexia nervosa. (AN) Folate and cobalamin of patients did not show signi®- and after nutritional rehabilitation (2) in a group of 11 patients. Plasma total-homocysteine in anorexia nervosa D Moyano et al 174 cant differences in comparison with the reference values. Nevertheless, tHcy measurement does not predict the They were in the reference range in patients except for one anorexic group since tHcy values of 63% patients over- case (folate: 3.2 nmol/L), who showed the highest tHcy lapped with reference values, but it proves rather useful to concentration (30 mmol/L). Creatinine concentration was detect the subclinical vitamin de®ciency in certain patients. signi®cantly higher in patients (P < 0.0001), although As regards related sulfur amino acids (Table 1), methio- within the reference range. Haemoglobin and MCV nine was signi®cantly increased (P < 0.02), probably owing showed normal values in all the patients. No signi®cant to muscle protein breakdown in order to cover energy correlation was found among tHcy, and folate demands and to preserve visceral proteins (Moyano et al, (r 7 0.255), cobalamin (r 7 0.096), cystine 1996). The rather mild elevation of this precursor amino (r 0.486), methionine (r 0.147), hemoglobin acid does not appear to directly determine increased tHcy (r 0.058), creatinine (r 0.472), MCV (r 0.090), and values, since no correlation was found between both amino BMI (r 7 0.090). However, there was a trend towards a acid concentrations. However, high methionine values negative correlation between tHcy and folate, and a posi- would determine high S-adenosylmethionine concentra- tive correlation between tHcy and creatinine. tions, that inhibit the enzymes of the remethylation pathway (methylene tetrahydrofolate reductase and betaine methyltransferase) and enhance the catabolism of Hcy Discussion through the transsulfuration pathway. If vitamin B6 tends Hematologic abnormalities (megaloblastic anemia) and=or to be de®cient in anorexia, activity of cystathionine b- neurologic dysfunction, which are the classical ®ndings synthase and cystathioninase would be impaired, and Hcy suggesting cobalamin and folate de®ciency (Carmel, 1989; concentration would increase. This might also explain Ueland et al, 1993), are apparently normal in anorexic cystine de®ciency, found in our patients and stated also patients. Recomended micronutrient needs stated 2 mg by other authors (Ghisol® et al, 1978; Schneede et al, cobalamin and 150±180 mg folate for female adolescents 1994). Vitamin B6 could not be measured in these group of (Story, 1992). Given the poor nutricional intake of anorexic anorexic patients. Although vitamin B6 de®ciency seems to girls, a vitamin de®ciency seems inevitable, especially that determine plasma tHcy concentration to a lesser extent of folate, which is the most important quantitatively. (Ubbink et al, 1993), further studies should be carried out Therefore, despite the normal vitamin values (Table 1), on this subject. Cystine depletion could also be in¯uenced stated also by other authors (Nussbaum, 1992), the inves- by increasing demands for the synthesis of glutathione, the tigation of subtle folate and=or cobalamin de®ciency by the substrate for the antioxidant enzymes (glutathione perox- assessment of their intracellular function re¯ected in idase and reductase) of the main free radical scavenging plasma tHcy concentration may be useful to apply the (Moyano et al, 1997). adequate nutritional rehabilitation. tHcy values in anorexic girls do not signi®cantly corre- Plasma tHcy, as well as methylmalonic acid in serum late with any of the biochemical magnitudes and Hcy and urine (Rasmussen et al, 1989), have been demonstrated determinants measured, vitamins, amino acids, creatinine, to be sensitive markers of intracellular folate and cobala- BMI, and hematologic indices. However, in patients there min de®ciency (Allen et al, 1990). tHcy was measured and was a trend towards a positive correlation with creatinine found signi®cantly high in anorexia patients. In fact, tHcy and a negative correlation with folate, also observed in increases abruptly with serum cobalamin concentration macrobiotic infants (Schneede et al, 1994). Creatinine below normal values ( < 130 pmol=L) (Ueland et al, values found in our anorexic patients appeared quite 1993), which did not occur in any of our patients, and homogenous compared with reference values, and although increases moderately when serum folate concentration is signi®cantly high, they are within the reference range. In subnormal ( < 5 nmol=L) (Ueland et al, 1993), which is the fact, we are dealing with an homogenous group of patients case of only one of our patients with the highest tHcy (similarly low muscular mass and high physical activity). concentration (30 mmol=L). Nevertheless, 21% of our anor- Creatine (and creatinine) formation is dependent on S- exic patients had plasma folate concentration in the low adenosylmethionine, so that for every molecule of creatine normal range (5±10 nmol=L) and their median tHcy one molecule of Hcy is formed (Refsum et al, 1997). (12.4 mmol=L (range: 6.3±30)) was above the reference However, S-adenosylmethionine is involved in many range, which is similar to other autors observations (Kang other methylation reactions, and plays an important role et al, 1987; Ueland et al, 1993). However, although 37% of in the complex control of Hcy metabolism (Fowler et al, patients showed cobalamin concentration in the low normal 1997). In fact, high S-adenosylmethionine might interfere range (150±270 pmol=L), their tHcy median was within the with folate availability for the remethylation of Hcy to references range (9.6 mmol=L (5.8±30)). The speci®c sig- methionine and thus, with intracellular folate function, ni®cance of the high tHcy concentration in relation to indirectly contributing to high Hcy values. Subnormal cobalamin and folate de®ciency is uncertain, since tHcy folate concentrations tend to negatively correlate with increases in both conditions (Ueland et al, 1993). Methyl- high tHcy values (Ueland et al, 1993; Ubbink et al, malonic acid should also be measured to discriminate 1993), which is not the case when there is a genetic between isolated cobalamin and=or folate de®ciencies. cause determining the hyperhomocysteinemia (Kang et al, However, methylmalonate excretion (measured by gas 1996). In fact, nutritional rehabilitation without speci®c chromatography) in 13 anorexic patients at diagnosis was vitamin supplementation corrected tHcy abnormal values in within the reference range ( < 7 mmol=mol creatinine) all the patients (Figure 2), con®rming their nutritional (data not shown). This suggests that intracellular folate origin. depletion may be preponderant in this anorexic group. As stated by Ueland et al (1993), folate undernutrition result- Conclusions ing in a serum folate content in the low-normal range is not suf®cient for optimal remethylation of Hcy, explaining the In summary, tHcy concentration appears signi®cantly hyperhomocysteinemia. increased in adolescents with anorexia nervosa most prob- Plasma total-homocysteine in anorexia nervosa D Moyano et al 175 ably owing to a subclinical folate de®ciency. This might be Moyano D, Vilaseca MA, Puig R, Moreno J, Artuch R & RamoÂn F (1996): caused by both, intracellular folate de®ciency (insuf®cient AminoaÂcidos plasmaÂticos y sus relaciones como expresioÂn del estado nutricional en la Anorexia Nerviosa. QuõÂm. ClõÂn. 15, 355. folate intake), and impaired folate availability (in¯uenced Moyano D, Sierra C, Brandi N, Artuch R, Mira A, Garcia-Tornel S & by high S-adenosylmethionine). Abnormal plasma tHcy Vilaseca MA (1997): Antioxidant status in anorexia nervosa. Int. J. Eat. values were completely corrected after nutritional rehabi- Dis. (in press). litation. Nussbaum MP (1992): Nutritional conditions. In: Textbook of Adolescent Medicine, ed. L Bralaw, pp 536±541. Philadelphia: Saunders Company WB. References Rasmussen K, Moelby L & Jensen MK (1989): Studies on Methylmalonic acid in humans. II. Relationship between concentrations in serum and Allen RH, Stbler SP, Savage DG & Lindenbaum J (1990): Diagnosis of urinary escretion, and the correlation between serum cobalamin and cobalamin de®ciency I: usefulness of serum methylmalonic acid and accumulation of methylmalonic acid. Clin. Chem. 35, 2277±2280. total homocysteine concentrations. Am. J. Hematol. 34, 90±98. Refsum H, Fiskerstrand T, Guttormsen AB & Ueland PM (1997): Assess- American Psychiatric Association (1994): Diagnostic and Statistical ment of homocysteine status. J. Inher. Metab. Dis. 20, 286±294. Manual of Mental Disorders (4th edn). Washington, DC: APA Press. Roediger WEW (1995): New views on the pathogenesis of Kwashiorkor: Carmel R (1989): Subtle vitamine B12 de®ciency. Biol. Clin. Hematol. 11, methionine and other amino acids. J. Pediatr. Gastroenterol. Nutr. 21, 3±11. 130±136. FarreÂ C, Velasco J, RamonÂ F. Evaluation of Olympius AU-510 analyser. J. Schebendach J & Nussbaum MP (1992): Nutrition management in ado- Autom Chem 1991; 13, 217±220. lescents with eating disorders. In: Adolescent Medicine: State of the Art Fisher M, Golden NH, Katzman DK, Kreipe RE, Rees J, Schebendach J, Reviews, eds. Nussbaum MP, Dwyer JT. Philadelphia: Handle & Belfus. Sigman G, Ammerman S, Hoberman HM (1995): Eating disorders 3, 541±557 in adolescents: a background paper. J. Adolesc. Health. 16, 420± Schneede J, Dagnelie PC, Staveren WAV, Vollset SE, Refsum H & 437. Ueland PM (1994): Methylmalonic acid and homocysteine in plasma Fowler B (1997): Disorders of homocysteine metabolism. J. Inher. Metab. as indicators of functional cobalamin de®ciency in infants on macro- Dis. 20, 270±285. biotic diets. Pediatr. Res. 36, 194±201. Ghisol® J, Carlet P, Ser N, Salvayre R, Thouvenot JP & Duole C (1978): Story M (1992): Nutritional requeriments during adolescence. In: Bralow Plasma free amino acids in normal children and in patients with L, ed. Textbook of Adolescent Medicine, ed. L Bralaw, pp 75±84. proteincaloric malnutrition: fasting and infection. Pediat. Res. 12, Philadelphia: Saunders Company WB. 912±917. Ubbink JB, Vermaak WJH, van der Merwe A & Becker PJ (1993): Kang S-S, Wong PWK & Norusis M (1987): Homocysteine due to folate Vitamin B-12, vitamin B-6, and folate nutritional status in men with de®ciency. Metabolism 36, 458±462. hyperhomocysteinemia. Am. J. Clin. Nutr. 57, 47±53. Kang S-S, Wong PWK (1996): Genetic and non genetic factors for Ueland PM, Refsum H, Stabler SP, Malinow MR, Andersson A & Allen moderate hyperhomocysteinemie. Atheroscerosis 119, 135±138. RH (1993): Total Homocysteine in plasma or serum: Methods and Madruga E, Marcos A, Varela P, MilaÂn C & Taracena del PinÄal B (1993): clinical applications. Clin. Chem. 39, 1764±1779. ParaÂmetros a utilizar en la evaluacioÂn nutricional de la anorexia Vilaseca MA, Moyano D, Ferrer I & Artuch R (1997): Total-Homocys- nerviosa. An. Esp. Pediatr. 68±71. teine in a pediatric patients. Clin. Chem. 43, 690±692.