www.elsevier.com/locate/sleep

Sleep Medicine 9 (2008) 388–392

Original article

Homocysteine in restless legs syndrome

Cornelius G. Bachmann a, Nicolas Guth a, Kathrin Helmschmied a,Victor W. Armstrong b, Walter Paulus a, Svenja Happe a,c,*

a Department of Clinical Neurophysiology, University of Gottingen, Germanyb Department of Clinical Chemistry, University of Gottingen, Germany

c Department of Clinical Neurophysiology, Klinikum Bremen-Ost, University of Gottingen, Zuricher Strasse 40, 28325 Bremen, Germany

Received 12 March 2007; received in revised form 19 June 2007; accepted 27 June 2007Available online 27 September 2007

Abstract

Background and purpose: Total plasma homocysteine (tHcy) may be a risk factor for vascular diseases and is associated with renalfailure or deficiency of vitamin B12 or folate. Recently, elevated tHcy concentrations were observed in patients with Parkinson’s dis-ease (PD), particularly those under levodopa treatment. Our objective was to determine whether changes in tHcy are also found inpatients with restless legs syndrome (RLS) in relation to levodopa treatment and whether folate and vitamins B6 and B12 play a rolein RLS.Methods: In a total of 228 subjects, tHcy and B vitamin status (vitamins B6 and B12, folate) were studied: 97 patients with idiopathicRLS (40 under levodopa therapy), 39 with PD (25 under levodopa therapy), and 92 healthy controls adjusted for age and gender.Results: No significant differences were observed in tHcy levels between RLS patients and controls or between the RLS groups with-out treatment or with levodopa or dopamine agonist treatment. Mean tHcy was significantly higher in PD patients (13.8 lmol/l)than in either RLS patients (11.7 lmol/l) or controls (11.0 lmol/l; p < 0.001). There was an inverse association between tHcyand vitamin B12 in each group.Conclusions: RLS and, in particular, levodopa treatment in RLS are not associated with hyperhomocysteinemia. Elevated tHcycould, however, be confirmed in PD patients.� 2008 Published by Elsevier B.V.

Keywords: Restless legs syndrome (RLS); Homocysteine; Folate; Vitamin B6; Vitamin B12

1. Introduction

Numerous studies have suggested that an elevatedplasma concentration of homocysteine is associated withan increased risk of cardiovascular disease [1]. Homocys-teine plasma concentrations depend on the status offolate and vitamins B6 and B12 [2]. Hyperhomocystein-emia has been linked with depression [3] and dementia

1389-9457/$ - see front matter � 2008 Published by Elsevier B.V.

doi:10.1016/j.sleep.2007.06.016

* Corresponding author. Address: Department of Clinical Neuro-physiology, Klinikum Bremen-Ost, University of Gottingen, ZuricherStrasse 40, 28325 Bremen, Germany. Tel.: +49 421 408 2370; fax: +49421 408 2375.

E-mail address: shappe@gwdg.de (S. Happe).

[4]. Additionally, homocysteine has been shown to betoxic to dopaminergic neurons [5] and glial cells [6] andto induce sural axonal degeneration [7]. In recent years,elevated plasma homocysteine levels have been observedin patients with Parkinson’s disease (PD), in particular inthose patients under treatment with levodopa [8,9].Levodopa is methylated by a series of reactions involvingB vitamins, which convert methionine to homocysteine.Therefore, treatment with levodopa may lead toincreased homocysteine and decreased B vitamin levels[10]. Patients with restless legs syndrome (RLS) are alsotreated with levodopa, and it might be speculated that inthis patient group homocysteine levels will be increased,similar to the situation reported for PD patients.

C.G. Bachmann et al. / Sleep Medicine 9 (2008) 388–392 389

The aim of this study was to investigate whether inRLS patients homocysteine, folate, and vitamins B6

and B12 levels are altered compared to healthy controls,and whether levodopa treatment is similarly associatedwith increased homocysteine and decreased folate andvitamin B6 and B12 levels compared to PD patients.

2. Patients and methods

2.1. Patients

All patients with RLS were recruited between 2003and 2005 in the specialized outpatient clinics for RLSand movement disorders of the Department of ClinicalNeurophysiology, University of Gottingen, Germany.The clinical diagnosis was made by experienced neurol-ogists according to the Essential Criteria of the Interna-tional Restless Legs Syndrome Study Group (IRLSSG)[11]. Blood investigations and clinical examinations wereperformed to exclude symptomatic forms of RLS. Sever-ity of RLS was evaluated using the International Rest-less Legs Syndrome severity scale (IRLS) [12].Altogether, 40 RLS patients under treatment with levo-dopa, 20 RLS patients under treatment with dopamineagonists (one patient received a combination therapyof 100 mg levodopa daily and a dopamine agonist),and 37 RLS patients without any RLS-specific therapywere included. We also investigated 92 healthy controlsrecruited from relatives accompanying the RLS patientsand from an advertisement in the hospital. In addition,we examined 39 patients with PD, 25 patients receivinglevodopa monotherapy, including six patients with levo-dopa and additional entacapone treatment; the othersdid not receive any dopaminergics. These patients wererecruited from the same specialized outpatient clinics.Disease severity in PD patients was assessed using theUnified Parkinson’s Disease Rating Scale (UPDRS)motor score [13] and the Hoehn and Yahr scale [14].Before participating in the study, all patients andhealthy controls gave informed consent. The local ethicscommittee of the Medical Faculty of the University ofGottingen, Germany, approved the protocol.

2.2. Blood tests

Whole blood samples, taken the morning after anovernight fast, were collected from all patients duringan outpatient visit at the Department of Clinical Neuro-physiology. Patients who were on regular vitamin sup-plementation were not allowed to take part in thisstudy. A minority of the included controls and patientssporadically took vitamins (Table 1). There is no generalfolate supplementation of the diet in Germany [15].Homocysteine and the B vitamins were analyzed imme-diately at the Department of Clinical Chemistry, Uni-versity of Gottingen, Germany.

Plasma total homocysteine, defined as the sum ofall homocysteine subfractions (including free and pro-tein-bound forms) in plasma, was determined by fluo-rescence polarization immunoassay on an AxSymimmunoanalyser (Abbott, Abbott Park, IL, USA).Serum concentrations of vitamin B12 and folate werequantified using binding assays on an Elecsys immu-noassay analyzer (Roche Diagnostics, Indianapolis,IN, USA). Serum concentrations of vitamin B6

were determined by fluorescence HPLC using theVitamin B6 HPLC Kit (Immundiagnostik, Bensheim,Germany).

2.3. Statistics

The v2 test (Fisher’s exact test, two-sided) was usedfor analysis of qualitative parameters.

Mean levels of homocysteine were compared betweenthree RLS groups as well as between RLS patients, PDpatients and healthy controls using analysis of variance(ANOVA) and the adjusted means procedure, adjustedfor age and gender. Since the values of the three vita-mins, folate, vitamin B6, and vitamin B12, as well as fer-ritin were not normally distributed, we log-transformedthem first and subsequently used ANOVA to calculateage- and gender-adjusted means for the different groups.Linear regression models were used to analyze the effectsof disease-specific factors (disease severity, disease dura-tion, treatment with levodopa and dopamine agonists)as well as disease-unspecific factors (body mass index[BMI], creatinine, urea, ferritin, vitamin supplementa-tion) on homocysteine levels stratified by type of disease.For this purpose, a linear regression model was used ineach group of participants. Correlation analysis wasdone using the Spearman rank test. Statistical analyseswere done using the STATA statistical package (version6.0).

3. Results

3.1. Clinical characteristics and demographics

There were no differences in age between the threeRLS groups or between RLS patients, PD patientsand healthy controls. No differences were observedbetween RLS patients and controls or between PDpatients and controls with regard to gender. However,there were fewer females in the PD group compared tothe RLS group. Disease duration in all RLS patientswas longer than in PD patients, particularly in thoseRLS patients who were on treatment with a dopamineagonist. In PD patients, the average daily levodopa dosewas higher (294.2 ± 315.7 mg) than that of RLS patients(91.8 ± 142.1 mg). The IRLS score was not differentbetween the three RLS groups. For further details seeTable 1.

Table 1Clinical and demographic data of all investigated patients with restless legs syndrome (RLS) (n = 97), Parkinson’s disease (PD) (n = 39), and healthycontrols (n = 92)

RLS patientswithouttreatment(n = 37)

RLS patientstreated withlevodopa(n = 40)

RLS patientstreated withdopamineagonist(n = 20)

p-Valuea All RLSpatients(n = 97)

All PDpatients(n = 39)

healthycontrolsubjects(n = 92)

p-Valueb

Age, years 59.4 ± 12.0 61.3 ± 11.7 59.3 ± 11.5 0.724 60.2 ± 11.7 64.3 ± 8.2 60.0 ± 11.56 0.096Female, n (%) 26 (70.3) 32 (80.0) 15 (75.0) 0.613 73 (75.3) 19 (48.7) 57 (62.0) 0.009Duration of disease,months, median

55 48 103 0.051 57 36 – 0.006

Levodopa dosage,mg (range)

– 216.3 ± 148.7(50–750)

5.0 ± 22.4(0–100)

0.001 91.8 ± 142.1 (0–750) 294.2 ± 315.7(0–1500)

– 0.001

RLS severity score,mean (range)

22.8 ± 7.0(10–39)

25.9 ± 5.5(13–37)

26.9 ± 9.4(7–37)

0.067 24.9 ± 7.2 (7–39) – – –

UPDRS motor score,mean

– – – – – 17.6 ± 11.0 – –

Hoehn & Yahr stage,median (range)

– – – – – 2.0 (1.0–3.0) – –

Vitaminsupplementation, n (%)

2 (5.4) 5 (12.5) 2 (10.0) 0.559 9 (9.3) 4 (10.8) 9 (9.8) 0.982

Body mass index 21.9 ± 2.8 23.2 ± 4.5 21.3 ± 2.7 0.103 22.3 ± 3.7 22.1 ± 3.2 23.0 ± 4.6 0.403Urea, mg/dL 15.1 ± 3.5 15.5 ± 3.6 16.4 ± 8.1 0.634 15.5 ± 4.8 17.3 ± 4.6 §§ 15.1 ± 4.3 0.042Creatinine, mg/dL 0.9 ± 0.2 0.9 ± 0.2 0.9 ± 0.3 0.693 0.9 ± 0.2 0.9 ± 0.2 0.9 ± 0.2 0.768

Values are presented as means ± SD unless otherwise specified.a p-Values for age and gender adjusted difference between the three RLS groups.b p-Values for age and gender adjusted difference between all RLS and PD patients and healthy controls.

390 C.G. Bachmann et al. / Sleep Medicine 9 (2008) 388–392

3.2. Homocysteine

Mean plasma homocysteine levels adjusted for ageand gender were significantly higher (p < 0.001) in PDpatients compared to all RLS patients and healthy con-trols. Homocysteine plasma concentrations did not dif-fer between RLS patients and healthy controls, norwas any difference found between the three RLS sub-groups (Table 2). Homocysteine levels in PD patientson treatment with levodopa alone (13.9 ± 5.8 lmol/L;n = 19) and in patients with additional entacapone treat-ment (16.6 ± 1.6 lmol/L; n = 6) did not significantly

Table 2Homocysteine, folate, vitamin B6, and vitamin B12 of all investigated patient(n = 39), and healthy controls (n = 92)

RLS patientswithouttreatment(n = 37)

RLS patientstreated withlevodopa(n = 40)

RLS patientstreated withdopamineagonist(n = 20)

Homocysteine,lmol/L

11.9 11.2 12.3

Folatea 2.27 2.37 2.35Vitamin B6

a 2.46 2.30 2.55Vitamin B12

a 5.85 5.60 5.86

Values are presented as means ± SD unless otherwise specified. p-Values wea Values were log-transformed due to non-normal distribution.b p-Values for age and gender adjusted difference between the three RLS gc p-Values for age and gender adjusted difference between all RLS and PD* p < 0.001 to controls and p = 0.04 to RLS patients.

differ from the concentrations in PD patients withouttreatment (14.3 ± 6.1 lmol/L; n = 14) (p = 0.477). Therewas no correlation between homocysteine levels andlevodopa dose, neither in RLS patients nor in PDpatients receiving levodopa. There was also no age- orgender-adjusted relation between homocysteine and fer-ritin levels in any of the groups (data not shown).

3.3. B vitamins

Compared to healthy controls and RLS patients,patients with PD had significantly lower age- and gen-

s with restless legs syndrome (RLS) (n = 97), Parkinson’s disease (PD)

p-Valueb All RLSpatients(n = 97)

All PDpatients(n = 39)

healthycontrolsubjects(n = 92)

p-Valuec

0.392 11.7 13.8* 11.0 <0.001

0.656 2.32 2.10 2.31 0.030.413 2.42 1.99 2.50 0.0020.639 5.89 5.86 6.01 0.146

re derived using ANOVA, adjusted for age and gender.

roups.patients and healthy controls.

C.G. Bachmann et al. / Sleep Medicine 9 (2008) 388–392 391

der-adjusted plasma concentrations of folate (p = 0.03)and vitamin B6 (p = 0.002), whereas vitamin B12 levelsdid not differ between the three groups (Table 2). Plasmaconcentrations of folate, vitamin B6, and vitamin B12

showed no significant differences among the three RLSpatient subgroups. In all RLS patients, the log-trans-formed values of the vitamins correspond to 10.14lg/L folate, 11.23 lg/L vitamin B6, and 359.6 ng/L vita-min B12, respectively.

3.4. Linear regression analysis

After adjustment for age and gender, there was nosignificant influence of disease-specific factors such asdisease severity, disease duration, treatment with levo-dopa and dopamine agonists on homocysteine plasmalevels in RLS patients or in PD patients. The disease-unspecific factors BMI and vitamin supplementationstatus also showed no significant influence on homocys-teine plasma levels in either patient group nor in healthycontrols. Creatinine and urea showed a significant directinfluence, and vitamin B12 a significant indirect influ-ence, on homocysteine plasma levels in each group ofparticipants. Folate and vitamin B6 also significantlyinfluenced homocysteine levels in PD patients, whereasin healthy controls only folate showed a significanteffect, and in RLS patients both folate and vitamin B6

failed to significantly influence homocysteine levels.

4. Discussion

This study shows that RLS and, in particular, levo-dopa treatment in RLS are not associated with hype-rhomocysteinemia, whereas we could confirm theoccurrence of hyperhomocysteinemia in PD patients.

Low plasma concentrations of the key cofactors ofhomocysteine metabolism, folate, vitamin B6, and vita-min B12 are known to be associated with hyperhomocy-steinemia [16,17] as we could confirm in the presentstudy. Only a few patients took vitamin supplementa-tion, and we do not believe that this had a relevant influ-ence on the results because the proportion (about 10%)was similar in each patient group and in healthycontrols.

In confirmation of previous findings [18], our datademonstrate a positive association between age andhomocysteine plasma concentration in each group. Ele-vated homocysteine plasma concentrations in the elderlymay be due to decreased activity of cystathionine syn-thetase or may result from an impaired renal function[19]. Creatinine and urea concentrations showed a directinfluence on homocysteine concentrations in eachgroup, underlining an impaired renal function as onereason for higher plasma homocysteine levels.

The prevalence of RLS is known to be twofold higherin women than in men [20]. In line with this, there were

more female RLS patients than male patients in thepresent study. Consequently, more female controls wererecruited to correspond with the RLS patients’ genderdistribution. Prior investigation of homocysteine plasmaconcentrations in women and men revealed lower homo-cysteine levels in women [21]. In support of this observa-tion, we adjusted our analysis not only for age but alsofor gender. Therefore, these known covariates on homo-cysteine plasma levels did not explain our results.

Previous investigations have shown that plasmahomocysteine concentrations in untreated patients withPD and healthy controls do not vary considerably [22],whereas initiation of levodopa therapy has been shownto elevate homocysteine plasma concentrations in previ-ously untreated PD patients [10,23]. Since there was nodifference in homocysteine levels between PD patientstreated with levodopa and PD patients not on levodopa,and since there was no correlation between levodopadose and homocysteine plasma levels, the observedhyperhomocysteinemia in PD patients cannot be solelyexplained by the amount of levodopa intake or theintake of the catechol-O-methyl-transferase (COMT)-inhibitor entacapone, as controversially discussed inthe literature [24]. Our RLS patients received an averagedaily dose of only 91 mg levodopa, and no significanthomocysteine plasma elevation was observed, suggest-ing that this dosage might be too low to affect homocys-teine metabolism. It may be possible that RLS patientswith higher levodopa dosages will have elevated homo-cysteine plasma levels similar to PD. However, thisneeds to be further investigated.

When translating preclinical data to a clinical appli-cation, specifically data which indicate the toxic poten-tial of homocysteine to the dopaminergic system in adose-dependent manner [5,6], it is tempting to speculatethat levodopa treatment may accelerate the advance ofPD via elevated homocysteine levels. However, similarto previous studies, we also could not find a significantinfluence of disease-specific factors, such as diseaseseverity and disease duration, on homocysteine levelsin PD patients [25].

In summary, this case-control study shows that RLSand, in particular, levodopa treatment in RLS is notassociated with hyperhomocysteinemia, whereas ele-vated levels of homocysteine could be confirmed forPD patients on treatment with a higher dose of levo-dopa. Therefore, screening of homocysteine levels andB vitamin supplementation are not necessary in RLSpatients when receiving relatively low levodopa dosescompared to PD patients.

Acknowledgement

The authors thank Klaus Berger MD MPH, Univer-sity of Munster, Germany, for helpful statistical advice.

392 C.G. Bachmann et al. / Sleep Medicine 9 (2008) 388–392

References

[1] Furie KL, Kelly PJ. Homocyst(e)ine and stroke. Semin Neurol2006;26:24–32.

[2] Zhu BT. On the mechanism of homocysteine pathophysiologyand pathogenesis: a unifying hypothesis. Histol Histopathol2002;17:1283–91.

[3] Chen CS, Tsai JC, Tsang HY, et al. Homocysteine levels,MTHFR C677T genotype, and MRI hyperintensities in late-onset major depressive disorder. Am J Geriatr Psychiatry2005;13:869–75.

[4] Quadri P, Fragiacomo C, Pezzati R, Zanda E, Tettamanti M,Lucca U. Homocysteine and B vitamins in mild cognitiveimpairment and dementia. Clin Chem Lab Med2005;43:1096–100.

[5] Lee ES, Chen H, Soliman KF, Charlton CG. Effects of homo-cysteine on the dopaminergic system and behavior in rodents.Neurotoxicology 2005;26:361–71.

[6] Maler JM, Wiltfang J, Huther G, et al. Vitamin B12 schutztAstrozyten vor Homocystein-induzierter Toxizitat. Der Nerve-narzt 2003;74(Suppl 2):A 0290. Abstract.

[7] Muller T, Renger K, Kuhn W. Levodopa-associated increase ofhomocysteine levels and sural axonal neurodegeneration. ArchNeurol 2004;61(5):657–60.

[8] Kuhn W, Roebroek R, Blom H, van Oppenraaij D, Muller T.Hyperhomocysteinaemia in Parkinson’s disease. J Neurol1998;245:811–2.

[9] Muller T, Woitalla D, Hauptmann B, Fowler B, Kuhn W.Decrease of methionine and S-adenosylmethionine and increaseof homocysteine in treated patients with Parkinson’s disease.Neurosci Lett 2001;308(27):54–6.

[10] O’Suilleabhain PE, Bottiglieri T, Dewey Jr RB, Sharma S, Diaz-Arrastia R. Modest increase in plasma homocysteine followslevodopa initiation in Parkinson’s disease. Mov Disord2004;19:1403–8.

[11] Allen R, Picchietti D, Hening W, Trenkwalder C, Walters A,Montplaisir J. Restless legs syndrome: diagnostic criteria, specialconsiderations, and epidemiology. A report from the restless legssyndrome diagnosis and epidemiology workshop at the NationalInstitute of Health. Sleep Med 2003;4:101–19.

[12] Walters AS, LeBrocq C, Dhar Aand the International RestlessLegs Syndrome Study Group. Validation of the internationalrestless legs syndrome study group rating scale for restless legssyndrome. Sleep Med 2003;4:121–32.

[13] Fahn S, Elton RLand members of the UPDRS development inParkinson’s disease. Unified Parkinson’s disease rating scale. In:

Fahn S, Marsden CD, Calne DB, Lieberman A, editors. Recentdevelopments in Parkinson’s disease. NJ: MacMillan Care Infor-mation, Florham Park; 1990. p. 153–63.

[14] Hoehn MM, Yahr MD. Parkinsonism: onset, progression andmortality. Neurology 1967;17:427–42.

[15] Koletzko B, Pietrzik K. Medical importance of dietary supple-mentation with folate. Dtsch Arztebla 2004;101:A 1670–81.

[16] Durand P, Prost M, Loreau N, Lussier-Cacan S, Blache D.Impaired homocysteine metabolism and atherothrombotic dis-ease. Lab Invest 2001;81:645–72.

[17] Gartler SM, Hornung SK, Motulsky AG. Effect of chronologicage on induction of cystathionine synthase, uroporphyrinogen Isynthase, and glucose-6-phosphate dehydrogenase activities inlymphocytes. Proc Natl Acad Sci USA. 1981;78:1916–9.

[18] Ganji V, Kafai MR. Third National Health and NutritionExamination Survey. Demographic, health, lifestyle, and bloodvitamin determinants of serum total homocysteine concentrationsin the third National Health and Nutrition Examination Survey.Am J Clin Nutr 2003;77:826–33.

[19] Norlund L, Grubb A, Flex G, et al. The increase of plasmahomocysteine concentrations with age is partly due to thedeterioration of renal function as determined by plasma cystatinC. Clin Chem Lab Med 1998;36:175–8.

[20] Berger K, Luedemann J, Trenkwalder C, John U, Kessler C. Sexand the risk of restless legs syndrome in the general population.Arch Intern Med 2004;164:196–202.

[21] Jacobsen DW, Gatautis VJ, Green R, et al. Rapid HPLCdetermination of total homocysteine and other thiols in serumand plasma: sex differences and correlation with cobalamin andfolate concentrations in healthy subjects. Clin Chem1994;40:873–81.

[22] Muller T, Werne B, Fowler B, Kuhn W. Nigral endothelialdysfunction, homocysteine, and Parkinson’s disease. Lancet1999;354:126–7.

[23] Postuma RB, Espay AJ, Zadikoff C, Suchowersky O, Martin WR,Lafontaine AL, et al. Vitamins and entacapone in levodopa-induced hyperhomocysteinemia: a randomized controlled study.Neurology 2006;27(66):1941–3.

[24] Zesiewicz TA, Wecker L, Sullivan KL, Merlin LR, Hauser RA.The controversy concerning plasma homocysteine in Parkinsondisease patients treated with levodopa alone or with entacapone:effects of vitamin status. Clin Neuropharmacol 2006;29:106–11.Review.

[25] O’Suilleabhain PE, Sung V, Hernandez C, et al. Elevated plasmahomocysteine level in patients with Parkinson’s disease: motor,affective, and cognitive associations. Arch Neurol 2004;61:865–8.