j ourna l homepage: www.e lsev ie r.com/ locate /bone
Bisphosphonates in phenytoin-induced bone disorder
Suruchi Khanna, Krishna K. Pillai, Divya Vohora ⁎Department of Pharmacology, Faculty of Pharmacy, Jamia Hamdard (Hamdard University), New Delhi 110062, India
Chronic administration of phenytoin (PHT) has been associated with bone loss. Bisphosphonates [alendronate(ALD), ibandronate (IBD) and risedronate (RSD)] are potential candidates to prevent PHT-induced bonedisorders, and the present study evaluated their effect on the antiepileptic efficacy of PHT. The PHT-induceddepletion in folic acid (FA), vitamin B6 and vitamin B12 results in hyperhomocysteinemia. The elevatedcirculating homocysteine (hcy) could be a risk indicator for micronutrient-deficiency-related osteoporosis viageneration of free radicals. Thus, an attempt was also made to unravel the PHT's and bisphosphonates' effecton hcy. Male mice received PHT (35 mg/kg, p.o.) for 90 days to induce bone loss. ALD, RSD and IBD wereadministered orally at doses 0.65 mg/kg, 0.33 mg/kg, and 0.17 mg/kg respectively, for prevention and 1.3 mg/kg, 0.65 mg/kg, and 0.33 mg/kg respectively, for treatment of PHT-induced bone loss. The bone loss wasconfirmed by bone mineral density (BMD) analysis and bone turnover markers. Serum levels of hcy and FAwere estimated along with hydrogen peroxide levels and total antioxidant capacity in order to assess theantioxidant profile of bisphosphonates. The induction of bone loss by PHT was marked by lowered BMD andaltered bone turnovers. ALD and RSD administration to PHT treated groups significantly reverted the bonyadverse effects. No such effects were observed with IBD. In the bisphosphonates treated groups, hcy levelswere statistically at par with the control group. PHT at 35 mg/kg, p.o. could compromise bone mass and thus,could be a model of bone demineralization in mice. The ALD, IBD and RSD have no pharmacodynamicinteraction when administered along with PHT at the experimental level. Thus, their usage in the man-agement of PHT-induced bone disease could be worthwhile if clinically approved.
Epilepsy and AEDs intricately modulate the bone microarchitec-ture and BMD, affecting bone strength. For more than four decades,antiepileptic drugs (AEDs) have been known to cause serious effectson bone mineral density (BMD) [1–3]. AED therapy causes multipleabnormalities in calcium and bone metabolism, varying fromincreased bone turnover without significant loss of cortical or trabec-ular bone to osteopenia/osteoporosis and to osteomalacic disorder.Gross malformations in the bones allied mainly, but not solely, withthe cytochrome P450 (CYP450)-inducing AEDs and these AEDs mayact as an add-on to risk factors (e.g. seizure-precipitated falls andtrauma and sedentary lifestyle) for fractures in epileptics [4].
Phenytoin (PHT) is a potent hepatic mixed-function oxidase(CYP450) inducer including CYP1A2, CYP2C9, CYP2C19 and CYP3A4,as well as glucuronyl transferases and epoxide hydrolase [5]. CYP450induction influences calcium–vitamin D axis by reducing bio-availablevitamin D. Lowered circulating calcium owing to hypovitaminosis Dresults in compensatory secondary hyperparathyroidism. The en-hanced serum parathormone levels increase the bone calcium
mobilization and consequent bone turnover. Other pathophysiolog-ical mechanisms involved in PHT-induced bone loss may includecalcitonin deficiency, vitamin K deficiency, deprived estrogen levels,interventions with circulating homocysteine (hcy) levels and inhib-itory effect on collagen synthesis in cultured bone [4].
With the advent of newer medications having less number ofside effects, PHT is still considered a first line drug to treat epilepsy(generalized tonic–clonic seizures and status epilepticus) [6,7]. Also,the AED-induced bone loss has been reported to be more pronouncedwith PHT intake [8,9]. PHT-induced bone loss may be an inevitableeffect but can be mitigated by supplementation of anti-osteoporoticmedications.
Bisphosphonates, synthetic analogues of pyrophosphate, bind tohydroxyapatite at sites of active bone remodeling. Second generationbisphosphonate: alendronate (ALD) and third generation bispho-sphonates: ibandronate (IBD) and risedronate (RSD) are leadingnitrogen-containing bisphosphonates for prevention and treatmentof bone diseases viz. Paget's disease, hypercalcemia of malignancy,in various experimental models of osteoporosis [10]. They are FDA-approved anti-osteoporotic agents and are potent inhibitors of oste-oclastic action [11].
During the course of anti-osteoporotic regimen (mainly onbisphosphonates, raloxifene, among others), if a patient is laterdiagnosed with epilepsy syndrome, it is yet to be documented which
anti-osteoporotic drugs should be administered along with AEDs. It isinteresting to note that the bisphosphonates have emerged as first-line treatment drugs for bone loss [12,13]. Therefore, it is necessary tocarry out research aimed at discovering novel bisphosphonates thatcan be prescribed alongside PHT.
5-Methyl-tetrahydrofolic acid generated from folic acid (FA) isessential for hcy metabolism. Increase in the circulating hcy with PHTintake may be due to interference with FA, vitamin B6 and vitaminB12 metabolism [14]. Hcy not only interferes with collagen cross-linking during the bone remodeling process but also shows positivecorrelation with free radical generation. Thus, hyperhomocysteinemiamay be among several potential mechanisms involved in pathogen-esis of bone malformations [4].
In view of increased prevalence of the AEDs-induced bonedisorders, and the recent proposal for the use of bisphosphonates inAED therapy [15,16]; we endeavored to determine by way of this invivo experiment whether their administration in PHT therapy willaffect PHT's antiepileptic efficacy. In addition, the present studyevaluated their effect on PHT-induced alterations in the biochemicalmarkers of bone turnover and parameters related to probablemechanisms mainly hyperhomocysteinemia leading to bone loss, incomparison with calcium and vitamin D3 (CVD) supplementation.
Material and methods
Experimental animals
All experiments were performed on adult male albino Swiss miceweighing 24–35 g. The animals were procured from the Central AnimalHouse, Jamia Hamdard (Hamdard University), New Delhi. Prior to thecommencement of the experiment, the animals were allowed toacclimate for 7 days; food and water were available ad libitum. Themice were kept in a room maintained at 23±2 °C and 55±15%humidity with a 12 h light–dark cycle. All animal procedures wereapproved by the ethics committee at our institution (Project no. 368,year: 2007) and performed in compliance with institutional guidelinesfor the care and handling of experimental animals.
Drugs and chemicals
The following drugs were used: phenytoin sodium (PHT) (Sigma-Aldrich, India); alendronate sodium (ALD) (Ranbaxy, India); risedro-nate sodium (RSD) (Fleming Laboratories Ltd., India); ibandronatesodium (IBD) (Sun Pharmaceuticals India Ltd., India); folic acid (FA)(Sigma-Aldrich, India); calcium and vitamin D3 (CVD) (Cipcal™, CiplaLtd., India); vitamin D3 (VD) (Calcikind™, Mankind Pharma Ltd.,India). All other chemicals used in this study were of analytical grade.
Oral administration of drugs
PHT, ALD, RSD, IBD and FA were dissolved in distilled water. CVDand VD were suspended in 1% aqueous carboxymethylcellulose. Allthe drugs were administered orally once daily for 7 days using afeeding needle.
Selection of doses
The doses of the bisphosphonates and CVD were calculated fromthe corresponding human doses (Table 1). PHT dose was calibrated onthe basis of the histopathological analysis of femur and the PHT'splasma concentration.
PHT at 70 mg/kg, p.o. (a converted dose from the previous study byNissen-Meyer et al. [17]) resulted in 30–40% mortality within 20 daysof drug administration. The probable reason could be higher plasmaconcentration at 70 mg/kg dose. Keeping the fact in view that thebone loss should be induced at clinically relevant drug concentration
therefore, the dose was reduced to 35 mg/kg. 35 mg/kg of PHT wasfound to be a suitable dose to induce bone loss within the therapeuticconcentration (10–20 μg/mL) i.e. the deteriorating effect on boneswas produced in the therapeutic range and not at the levels (N30 μg/mL) at which the toxicity signs begin to appear. The blood sampleswere withdrawn randomly to monitor the plasma concentration ofPHT. The plasma PHT levels were well within the therapeutic concen-tration when administered alone and along with bisphosphonates.
Experimental protocol
The experimental protocol was divided into the following parts
1. Preventive treatment: In this experiment, eleven groups of eightmice each were administered drugs once daily for the durationof three months: control group (0.5% CMC, 2 mL/kg); PHT (35 mg/kg); ALD (0.65 mg/kg), RSD (0.33 mg/kg), IBD (0.17 mg/kg), FA(0.13 mg/kg), PHT (35 mg/kg)+ALD (0.65 mg/kg), PHT (35 mg/kg)+RSD (0.33 mg/kg), PHT (35 mg/kg)+IBD (0.17 mg/kg),PHT (35 mg/kg)+FA (0.13 mg/kg), and PHT (35 mg/kg)+CVD(130 mg/kg+65 IU).
2. Therapeutic treatment: In this experiment, fifty four mice wereadministered PHT (35 mg/kg) once daily for the duration of threemonths. At the end of three months, the mice were divided intosix groups and administered the bisphosphonates, FA, CVD oncedaily along with PHT for a period of one month. The groups wereas follows: PHT (35 mg/kg); ALD (1.3 mg/kg); RSD (0.65 mg/kg);IBD (0.33 mg/kg); PHT (35 mg/kg)+ALD (1.3 mg/kg); PHT(35 mg/kg)+RSD (0.65 mg/kg); PHT (35 mg/kg)+IBD (0.33 mg/kg); PHT (35 mg/kg)+FA (0.13 mg/kg); PHT (35 mg/kg)+CVDD[CVD (130 mg/kg+65 IU)+VD (195 IU)].
At the end of each treatment, urine collection was done for 24 h.Themicewere fasted overnight prior to collection of blood samplesfrom tail vein for biochemical estimations. Immediately after bloodcollection the mice were euthanized for collection of brain tissue(for biochemical estimations) and femora (to measure BMD).
3. Electroconvulsive group: In order to evaluate the threshold forelectroconvulsions, four groups of mice each containing 10 mice pergroup, were administered the drugs as per therapeutic treatmentgroup.
Electroconvulsive threshold (maximal electroshock seizure threshold test)Electroconvulsions were produced by Electroconvulsive Treat-
ment Unit (UGO BASILE, Italy). The pro-/anti-convulsant potentialwas evaluated in the model of maximal electroshock seizurethreshold (MEST) test [18]. An alternating current (50 Hz, 0.2 s) wasdelivered via ear-clip electrodes. Tonic hindlimb extension (the hind
Table 2Effect of chronic administration of bisphosphonates alone and in combination with PHTin electroshock-induced seizures.
Treatment (mg/kg) CS50 (mA)
Vehicle 14.16±0.11ALD (0.65) 13.85±0.38ALD (1.3) 13.70±0.18RSD (0.33) 14.24±0.35RSD (0.65) 14.40±0.42IBD (0.17) 13.94±0.97IBD (0.33) 13.87±0.26PHT No hind limb extensionPHT+bisphosphonates No hind limb extension
Results are expressed as mean±SEM; CS50: a current strength inducing tonic–clonicconvulsions in 50% of tested mice.
599S. Khanna et al. / Bone 48 (2011) 597–606
limbs of animals outstretched 180° to the plane of the body axis) wastaken as the endpoint. The bisphosphonates were tested for theirpro-/anti-convulsant potential by determining their ability to protect50% of animals against the maximal electroshock-induced tonichindlimb extension and expressed as respective current strength. Todetermine the value of median current strength (CS50 in mA), at leastfour groups of mice (10 animals per group) were challenged withcurrents of various intensities. Then, a current intensity–effect curvewas constructed, according to a log-probit method by Litchfield andWilcoxon, [19] from which CS50 in mA was estimated.
To calculate CS50, per se groups of bisphosphonates were subjectedto various current intensities ranging from 10 to 20 mA at theirpeak plasma time intervals. In order to evaluate any possible variationin PHT's antiepileptic potential with bisphosphonates, PHT aloneand in combination with bisphosphonates were subjected to variouscurrent intensities ranging from 10 to 50 mA. The animals wereadministered bisphosphonates in such a way that their peak plasmalevels (approx 1 h)matched the peakplasma levels of PHT (approx 2 h).
Bone tissue sample preparation and assays performed
The left femora were cleaned of soft tissues and then frozen at−20 °C. Before the dual-energy X-ray absorptiometry (DEXA)examination for BMD analysis, they were first defrosted for 30 min.The femora were scanned by DEXA using a Hologic (QDR 4500, USA).
The right femora were dissected out and the surrounding musclesand tissues were removed. The femora wereweighed individually andhomogenized with 10 volumes of 10 mM triethanolamine buffer (pH7.5). The homogenate was stirred for 1.5 h at 4 °C and centrifuged. Theextraction procedure was repeated twice and aliquots of the boneextracts were used for the determination of the activities of alkalinephosphatase (ALP) and tartrate resistant acid phosphatase (TRAP).The insoluble pellets were hydrolyzed with 6 N HCl at 105 °C for 24 hand analyzed for hydroxyproline (HxP).
ALP activity was determined using commercial kit (SPAN Diag-nostics, India). TRAP activity was estimated by the method ofTenniswood et al. [20]. The protein estimation was done by themethod of Lowry et al. [21]. The amount of HxP was determinedaccording to the method of the chloramine-T oxidation procedure ofStegemann [22].
Serum preparation and assays performed
Serum was separated by centrifugation for 10 min at 3000 rpm.Serum samples were stored at −20 °C until analysis was carried out.Serum hcy levels were measured by using commercially available kitbased on the principle of fluorescence polarization immunoassay(FPIA) method on the AxSYM System analyzer (Abbott Laboratories,Abbott Park, IL). FA levels were estimated by using commerciallyavailable Solid Phase No Boil Assay Kit® (Seimens Medical SolutionDiagnostics, Los Angeles, USA). Total antioxidant capacity (TAC) wasestimated by the Koracevic et al. [23].
Urine collection and assays performed
Urine collection was done for 24 h in a clean flask placed on ice.Hydrogen peroxide (HP) determination was done by using commer-cially available Hydrogen Peroxide Assay Kit® (Cayman ChemicalCompany, Ann Arbor, MI)
Statistical analysis
Data analysis was carried out by using Graphpad Prism 3.0(Graphpad software; San Diego, CA). CS50 values with their respective95% confidence limits were estimated using computer log-probitanalysis according to Litchfield andWilcoxon [19]. Subsequently, SEM
values were calculated on the basis of confidence limits, number ofanimals and slope function obtained directly from log-probit analysisaccording to Litchfield and Wilcoxon [19]. The statistical evaluationof respective CS50 vs. control values was performed with analysisof variance (ANOVA) followed by post-hoc Bonferroni's test. Theother parameters mentioned earlier were expressed as mean±SEM.Groups of data were compared with ANOVA followed by Tukey–Kramer multiple comparison tests. Values were considered statisti-cally significant at Pb0.05.
Results
Electroconvulsive threshold test
CS50 values after treatment with the bisphosphonates alone areshown in Table 2. Bisphosphonates at both doses did not affect seizurethreshold. No significant change in CS50 values was observed ascompared to control group. PHT at 35 mg/kg afforded 100% protectionagainst MEST as evidenced by complete abolition of the tonic ex-tension phase. Combination groups of PHT and bisphosphonates (atboth the doses) were unable to produce any significant change in thetonic extension phase observed during seizure activity and elicitedsimilar responses as those with PHT per se. Bisphosphonates did notaffect the antiepileptic activity of PHT.
BMD analysis
The changes in BMD of the left femur determined by DEXA areshown in Fig. 1. There was a significant lowering in BMD levels of PHTgroup as compared with the control group (pb0.001). Both ALD andRSD, whether given concurrently or post PHT treatment substantiallyincreased the BMD levels as compared with the PHT group (pb0.001).Furthermore, the BMD level in the PHT+RSD group was significantlyhigher than that in the PHT+ALD group (pb0.01). IBD neitherprevented nor treated the PHT induced decrease in BMD (pN0.05).Both ALD and RSD were more effective than CVD and CVDD inpreventing the decline of BMD levels. FA treatment was ineffectivein preventing BMD reduction in PHT administered (preventive andtreatment) groups.
Bone turnover markers
a) ALP activityThe differences in the femoral ALP activity between the groups areshown in Table 3. Long-term treatment with PHT significantlyattenuated ALP activity in the femoral bones. ALD, IBD and RSDper se groups at both doses did not show any significant change inthe femoral ALP activity as compared to control group (pN0.05).ALD, RSD, CVD and CVDD prevented the diminutive effect of PHT
Fig. 1. Effects of PHT, bisphosphonates and their combination on femoral bone mineral density (BMD). Values are represented as mean±SEM; ***Pb0.001 versus control;###Pb0.001 versus PHT, nsPN0.05 versus PHT; $$Pb0.01 versus PHT+CVD/CVDD; $$$Pb0.001 versus PHT+CVD/CVDD.
600 S. Khanna et al. / Bone 48 (2011) 597–606
significantly (pb0.001). In addition, the concomitant administra-tion of PHT and FA reversed the decreased bone ALP levels in mice(pb0.05).
b) TRAP activityA significant elevation in TRAP levels was observed in the groupsof mice treated with PHT (Table 3). ALD, RSD, FA, CVD and CVDDreversed the elevated TRAP levels significantly in PHT treated(preventive and treatment) groups. IBD administration did notreverse elevated TRAP activity. The concomitant treatment of FAwith PHT and the post treatment of FA in mice treated with PHTwere found to slightly lower the elevated TRAP activity (pb0.05).
c) HxP concentrationTable 3 summarizes the effect of bisphosphonates and calcium andvitamin D supplements on HxP content. Femora of PHT-treatedgroups contained less HxP/mg of bone as compared to the vehiclecontrol group (pb0.001). In the groups receiving bisphosphonates,CVD and CVDD along with the PHT, HxP levels were elevated ascompared to PHT groups (pb0.001). FA and IBD did not alter HxPlevels when compared with PHT-administered mice (pN0.05).
TAC and HP measurement
Compared with the control group, the TAC was depleted in PHTtreated groups. As shown in Table 4, bisphosphonates, CVD and CVDDgroups statistically altered the TAC levels (pb0.001). Treatment withFA in PHT-administeredmicewas found to slightly enhance TAC levelsas compared to the mice treated with PHT (pb0.05).
Table 4 depicts the effect of drug treatments on urinary HP levels.HP levels were found to be considerably enhanced in PHT groups incomparison to the group administered with normal saline (pb0.001).Bisphosphonates, FA, CVD and CVDD treated groups had a significanteffect in lowering the elevated HP levels in comparison to the grouptreated with PHT (pb0.001).
Hcy and FA levels
The changes in the serum hcy levels between the groups arerepresented in Table 5. The serum hcy levels increased significantlywith PHT treatment and the anti-resorptive drugs (ALD, RSD and IBD)approximately lowered the elevated hcy to the baseline level.Bisphosphonates alone did not alter the serum hcy levels. CVD andCVDD administration statistically lowered the hcy levels in PHT-treated groups (pb0.05). The combined administration of FA with
PHT almost completely reversed the elevated hcy levels in PHTtreated groups. The results of PHT+ALD/RSD groups were at par withPHT+FA group.
Mice treated with PHT had decreased serum FA levels, whichwere significantly different from control group (pb0.01) as shown inTable 5. No significant differences were noted with FA levels betweenbisphosphonates per se and combinational groups of PHT and ALD/RSD/IBD/CVD/CVDD groups (pN0.05). However, the administrationof FA along with PHT-treated mice normalized the lowering effectof PHT on serum FA levels (pb0.001) in comparison to PHT-treatedmice.
Discussion
The study provided first experimental evidence on bisphospho-nates' effect on seizures and antiepileptic efficacy of PHT as well as onPHT-induced metabolic bone loss.
Bisphosphonates' effect on seizures and antiepileptic efficacy of PHT
Results presented in our study showed that bisphosphonates (ALD,RSD and IBD) at both doses revealed no appreciable changes in theCS50 as compared to the control group suggesting that the bispho-sphonates (ALD, RSD and IBD) have no inherent pro-/anti-convulsantproperty. Concurrent administration of PHT and bisphosphonates(ALD, RSD and IBD) elicited similar responses as those with PHT per se.This clearly indicates that bisphosphonates (ALD, RSD and IBD) didnot alter the electroconvulsive threshold of PHT in MEST test(Table 2). This rules out any possible interaction of bisphosphonateswith the PHT at the pharmacodynamic level on long-term therapy.
Bisphosphonates in PHT-induced metabolic bone loss
The prolong therapy with PHT has been associated with changesin bone turnover cycle [17,24]. At the end of a span of 30, 60 and90 days, the left femur and lumbar vertebrae (L2–L4) were dissectedout from the mice selected randomly in order to detect the changesrelated to bone loss through histopathological analysis (Figs. 2a–f).The femur dissected after 90 days showed the prominent thinningof bone matrix; an enhanced osteoclastic activity; and prominentruffled border (Fig. 2d). This is in line with the studies performed byValimaiki et al. [25] and Moro-Alvarez et al. [26] revealing that thePHT affects cortical bones more profoundly than the trabecular bones.
Table 4Effects of combinations of bisphosphonates with PHT on TAC and HP levels.
Groups (n=8)
Control PHT ALD IBD RSD FA PHT+FA PHT+ALD PHT+IBD PHT+RSD PHT+CVD PHT+CVDD
HxP (μg of HxP/mg of bone) [preventive treatment]156.95±9.188 104.64±7.856*** 147.86±7.364 147.13±7.803 151.06±7.061 148.70±6.027 113.45±7.013ns 156.67±6.794### 103.27±7.990ns 156.02±8.743### 156.18±6.837### –
HxP (μg of HxP/mg of bone) [therapeutic treatment]156.95±9.188 90.67±8.454*** 158.62±11.612 156.44±6.149 160.68±9.219 – 92.39±8.164ns 153.74±8.721### 105.26±10.056ns 151.61±10.051### – 152.68±7.900###
Values are represented as mean±SEM; n = number of animals; ***pb0.001 versus control; #pb0.05 versus PHT, ###pb0.001 versus PHT, nspN0.05 versus PHT.
601S.K
hannaet
al./Bone
48(2011)
597–606
Table5
Effectsof
bisp
hosp
hona
teson
PHT-indu
cedch
ange
sin
Hcy
andFA
leve
ls.
Group
s(n
=8)
Control
PHT
ALD
IBD
RSD
FAPH
T+
FAPH
T+
ALD
PHT+
IBD
PHT+
RSD
PHT+
CVD
PHT+
CVDD
Hcy
(μmol/L)[preve
ntivetrea
tmen
t]8.68
±0.80
119
.86±
2.04
8***
9.38
±0.75
58.59
±0.97
39.28
±1.00
18.97
±1.27
49.94
±0.62
5###
8.96
±1.12
1###
10.40±
0.84
8###
8.73
±0.94
9###
14.25±
1.51
5#–
Hcy
(μmol/L)[the
rape
utic
trea
tmen
t]8.68
±0.80
124
.75±
1.19
4***
8.32
±0.63
48.16
±0.57
98.58
±1.09
1–
9.93
±0.90
8###
10.79±
0.96
7###
12.01±
0.79
6###
10.85±
0.57
6###
–20
.27±
1.42
3#
FA(n
g/mL)
[preve
ntivetrea
tmen
t]5.64
±0.14
44.40
±0.13
0**
6.04
±0.40
65.46
±0.16
35.34
±0.18
65.74
±0.18
15.60
±0.11
4##
4.16
±0.14
4ns
4.32
±0.15
9ns
4.52
±0.22
2ns
4.48
±0.13
2ns
–
FA(n
g/mL)
[the
rape
utic
trea
tmen
t]5.64
±0.14
43.90
±0.07
1***
5.50
±0.23
05.66
±0.21
65.46
±0.22
5–
5.88
±0.08
6###
3.76
±0.09
3ns
3.68
±0.08
6ns
3.88
±0.15
9ns
–3.86
±0.15
0ns
Value
sarerepresen
tedas
mea
n±
SEM;n=
numbe
rof
anim
als;
**pb0.01
versus
control,***p
b0.00
1ve
rsus
control;
#pb0.05
versus
PHT,
##pb0.01
versus
PHT,
###pb0.00
1ve
rsus
PHT;
ns p
N0.05
versus
PHT.
602 S. Khanna et al. / Bone 48 (2011) 597–606
And, since the femur is mainly composed of cortical bone, the femurwas chosen to detect the bony changes induced by PHT. Thehistopathological changes observed with PHT were more pronouncedwith femur rather than lumbar vertebrae (L2–L4). To check whetheron extending the duration of PHT therapy, histopathological changesare observed in L2–L4 or not, as observed in other studies [26–29],lumbar vertebrae (L2–L4) of PHT-treated mice were dissected atthe end of 120 days. No histopathological changes were, however,observed in comparison with the control group vertebrae even at theend of 120 days.
BMD analysisBone loss was further confirmed by BMD analysis using DEXA
technique. The influence of PHT on femur was consistent with theprior studies documenting lowering of BMD (Fig. 1) [24,25,27]. Theseresults support that PHT at 35 mg/kg; p.o. could compromise bonemass and thus, be a model of bone demineralization in mice.
The efficacy of bisphosphonates in reducing vertebral and non-vertebral fractures by increasing bonemass has been demonstrated inlarge-scale clinical trials [30–35]. In the Fracture Intervention Trial(FIT) studies, ALD was found to reduce vertebral and non-vertebralfractures [30]. Similar results were observed in Vertebral Efficacy withRisedronate Therapy (VERT) trials in which RSD significantly reducedvertebral and non-vertebral fractures [32]. This is in concert with ourfindings reporting that ALD and RSD have significantly reverted thelowered BMD values of femora in PHT-treated mice (Fig. 1). Whetheradministered prophylactically or therapeutically, these agents accel-erate bone healing.
The Oral Ibandronate Osteoporosis Vertebral Fracture Trial inNorth America and Europe (BONE) study reported that IBD is moreefficacious for vertebral fractures than non-vertebral ones [33]. In ourstudy also, IBD was inefficacious in preventing femoral bone loss inthe PHT-treated mice. This finding is supported by an earlier reportstating the lack of evidence for its efficacy in hip and non-vertebralfracture reduction [33]. This may explain partly that IBDmay not be aseffective as ALD and RSD in preventing femoral bone loss. The positiveresults of BMD of CVD and CVDD treated mice were in line with thetrials done with 1000 mg calcium plus vitamin D3 400 IU [36,37]. Incomparison to CVD/CVDD, ALD and RSD statistically altered the BMDvalues, the alteration being more pronounced with RSD (pb0.01 forALD and pb0.001 for RSD).
PHT's multifactorial mechanisms of bone loss ultimately lead toosteoclastogenesis. It also stimulates osteoclast differentiation factor/receptor activator nuclear factor-kappa B ligand (RANKL) mRNAexpression directly and induces osteoclastogenesis [38]. Bispho-sphonates selectively disrupt osteoclast activity [39] and our studyfound them to be more effective in elevating BMD than CVD/CVDD.Further, Jowsey and co-workers reported that calcium and vitamin Dsupplementation cannot be expected to increase bonemass but mightslow further loss of mineralized tissue by suppressing bone resorption[40]. The highly selective localization, bone retention and anti-osteoclastic property make the bisphosphonates, the drug of choice.
Bone markersBone markers can predict bone loss independent of BMD; the
resorption markers being better indicator than formation markers[41]. Uncoupling of bone resorption and formation markers results inhigh bone turnover and bone loss. To enhance the specificity of bonemarkers, the analysis was done on femoral bone.
Markers of bone formation assess either osteoblastic syntheticactivity or post release metabolism of procollagen. ALP produced bythe osteoblasts, is an excellent indicator of early osteoblast differen-tiation as its level indicates the bone metabolic activity [42]. PHTadministration has been found to decrease ALP in the femur and thisindicates lower bone formation during bone remodeling andenhanced bone loss (Table 3). Anti-osteoporotic agents (ALD, RSD,
Fig. 2. Histology of femur (a–d) (40×): b: PHT treatment for 30 days showed no change in the osteoclastic activity on the border, no disruption of Harvesian system and no effect onmatrix; c: PHT treatment for 60 days showed slight thinning of bone matrix, slightly ruffled border and increased osteoclastic activity; d: PHT treatment for 120 days showedprominent thinning of bone matrix (lighter area), prominent ruffled border and enhanced osteoclastic activity. Histology of lumbar (e, f) (40×): f: PHT treatment for 120 daysshowed no change in the osteoclastic activity on the border, no disruption of Harvesian system and no effect on matrix.
603S. Khanna et al. / Bone 48 (2011) 597–606
CVD, and CVDD) enhanced the bone formation, observed as elevationin the ALP levels, whether given prophylactically or therapeutically.A slight elevation in ALP level (Pb0.05) was observed with IBD as apreventive agent. Our study is consistent with the previous studydemonstrating a positive correlation of ALP and the bisphosphonates[43]. This correlation could be due to bisphosphonates' anaboliceffect on osteoblasts via stimulating basic fibroblast growth factor(bFGF) production and up-regulating bone morphogenetic protein-2(BMP-2) gene expression [44,45]. Also, bisphosphonates may preventosteoblast apoptosis and indirectly contribute to the relative increasein cell number and activity [46]. Thus, this explains that enhancedosteoblastic activity by bisphosphonates could have increased ALPactivity, marker of bone formation.
Resorption markers reflect osteoclast activity and/or collagendegradation. HxP, a modified amino acid of collagen, is the product ofposttranslational hydroxylation of integral proline residues of type 1collagen [42]. Urinary HxP, though clinically useful, has several lim-itations such as: lack of tissue specificity; its origin during inflamma-tory conditions; its extensive metabolic degradation before urinary
excretion; an obstructive sensitivity to diet [47,48]. Thus, it wasplanned to assess HxP in the bone. Decrease in HxP content wasobserved in the femora of PHT-treated mice (Table 3). This could berelated to the fact that PHT might have modified the collagen naturewhich might have altered the stabilizing cross-links, which in turnhas reported to affect bone's mechanical properties [49]. In our study,bisphosphonates (except IBD) and calcium and vitamin D supple-ments significantly reverted the decreased HxP levels. Hence, it canbe said that the bisphosphonates improve the bone's mechanicalstrength.
Another, bone resorption marker, TRAP is secreted at the ruffledborder by osteoclasts, dephosphorylates osteopontin (bone matrixprotein, TRAP substrate) and thus, results in bone resorption [42].TRAP levels were found to be increased in the femur of PHT-treatedmice suggesting the enhanced bone loss with PHT administration(Table 3). The reduction in TRAP activity as observed with ALD andRSD administration was comparable with the CVD and CVDD.Bisphosphonates have been reported to affect osteoclastogenesisand induce apoptosis in osteoclasts directly or indirectly [11,38,39].
The results of our study provide a line of evidence similar to that inthe study by Nishikawa and co-workers in which bisphosphonatesinhibited osteoclasts-related TRAP release [50].
Thus, it could be inferred that the changes observed in the data ofbone turnover markers are similar to the changes observed in BMD.
Bisphosphonates and PHT-induced hyperhomocysteinemia
Hcy homoeostasis is maintained by the concerted processes(remethylation, transmethylation and trans-sulphuration); involvingFA (cofactor in one-carbon transfer reaction), vitamin B12 (coenzymein the methyl transfer from 5-methyltetrahydrofolate to hcy) andvitamin B6 (cofactor in the trans-sulphuration pathway) [51]. Def-iciency of either of the cofactors or coenzyme resulting from PHTadministration is likely to perturb hcy homeostasis and this con-sequently leads to accumulation of hcy in the blood [52]. In our study,PHT too induced Hhcy (Table 5). Our results are consistent withthe reported clinical as well as experimental studies documenting thedeficit in serum FA levels on PHT treatment (Table 5) [53–55]. Thepossible explanations to the decreased FA levels are [53,56]: 1) PHTmay interfere with the intestinal absorption of FA (by increasingintestinal pH); or 2) it may induce FA catabolizing enzyme (CYP450);or 3) it may interfere with the metabolism of FA coenzymes.
We have found a significant elevation in serum hcy levels fol-lowing PHT treatment. FA administration was found to lower theaugmented hcy both prophylactically and therapeutically (Table 5)and thus, supporting the already proven anti-hyperhomocysteinemicpotential of FA [57]. The combination of bisphosphonates and PHTdid not alter the decreased serum FA levels in the PHT-treatedgroups (Table 5). Also, CVD and CVDD treatment showed no improve-ment in FA levels in the PHT-treated groups. However, bispho-sphonates completely reversed the elevated hcy levels in the PHT-induced groups. Significant lowering of hcy levels were observed withCVD and CVDD administration but to a lesser extent in comparison tobisphosphonates. The probable explanation could be substantiated as:first, bisphosphonates did not interfere with B vitamin dependentremethylation pathway of hcy. Second, lowering of the hcy levels bybisphosphonates could be due to their antioxidant profile (Fig. 3).
Fig. 3. Probable mechanism of hcy lowering effect of bisphosphonates. The autoxidation of hosteoclast activity and prevent the release of TRAP. They have been found to quench HP an
Bisphosphonates in hyperhomocysteinemia-related bone loss
PHT-induced Hhcy may predispose epileptics to micronutrient-deficiency-related bone loss [58,59]. Elevated serum hcy has beenfound to disrupt redox homeostasis by enhanced free radical produc-tion. The autoxidation of hcy to homocysteine liberates HP, which canform oxidants (hydroxyl radicals) [60]. The free radical generationowing to HP accumulation can trigger a cascade of events contributingto pathological changes in bone by suppressing the formation (inhibitosteoblastic differentiation) and stimulating the resorption (stimulateosteoclast differentiation) of bone [61,62]. We found that the bis-phosphonates quenched HP thereby, reducing the HP concentration(Table 4).
The hcy's pro-oxidant potential was further confirmed by de-creased serum TAC levels as it measures the capacity of the serum toinhibit the thiobarbituric acid reactive substance (TBARS) production[23]. PHT-induced Hhcy lowered serum TAC levels. The increase inserum TAC levels observed in bisphosphonate treated groups, furtherjustified the antioxidant properties of bisphosphonates (Table 4).
TRAP catalyzes free radical generation at the ruffled border andenhances bone resorption [63]. These free radicals may also result inbreakdown of bone collagen (HxP). As explained earlier, ALD, RSD,CVD and CVDD increased HxP levels and decreased TRAP activity inthe femur. These anti-osteoporotic agents oppose the TRAP's effectand thus, inhibit orthophosphate monoesters' hydrolysis and freeradical-induced bone loss.
Overall, the results of the present study add to the accumulatingevidence suggestive of induction of bone loss with PHT on chronictherapy. The alterations in hcy pathway and the free radicalgeneration by PHT could be reversed with bisphosphonate adminis-tration. Both the second generation (ALD) and third generation (RSD)bisphosphonates are efficacious to revert PHT's bony effects. Althoughthe precise mechanism is yet to be elucidated, it can be hypothesizedthat the use of ALD and RSD could be safe and potentially useful in thePHT-induced bone loss. It is clear from the present study that such anaddition (ALD, RSD and IBD) will not influence the PHT's antiepilepticefficacy. Further studies are absolutely necessary to delineate anyconclusion regarding the chronic incorporation of bisphosphonates inPHT induced bone abnormalities.
cy to homocysteine liberates HP, which can form free radicals. Bisphosphonates disruptd free radicals.
The authors are thankful for the financial assistance from UGC,New Delhi. We also acknowledge the help of Prof. A.C. Ammini, Head,Dept of Endocrinology and Prof. Y. K. Gupta, Head, Dept. Pharmacol-ogy, AIIMS for extending necessary facilities.
References
[1] Kruse R. Osteopathies in antiepileptic long-term therapy (preliminary report).Monatsschr Kinderheilkd 1968;116:378–81.
[2] Christiansen C, Rodbro P, Lund M. Incidence of anticonvulsant osteomalacia andeffect of vitamin D: controlled therapeutic trial. Br Med J 1973;4:695–701.
[3] Lee RH, Lyles KW, Emeric CC. A review of the effect of anticonvulsant medicationson bone mineral density and fracture risk. Am J Geriatr Pharmacother 2010;8:34–46.
[4] Khanna S, Pillai KK, Vohora D. Insights into liaison between antiepileptic drugs andbone. Drug Discov Today 2009;14:428–35.
[5] Patsalos PN, Froscher W, Pisani F, Van Rijn CM. The importance of druginteractions in epilepsy therapy. Epilepsia 2002;43:365–85.
[6] Cherian A, Thomas SV. Status epilepticus. Ann Indian Acad Neurol 2009;12:140–53.
[7] McNamara JO. Pharmacotherapy of the epilepsies. In: Goodman LS, Gilman A,Brunton LL, Parker KL, editors. Goodman and Gilman's manual of pharmacologyand therapeutics. New York: McGraw-Hill Medical; 2008. p. 333–4.
[8] AlisonMP, Martha JM. Epilepsy and bone health in adults. Epilepsy Beh 2004;5(2):24–9.
[9] Valimaki MJ, Tiihonen M, Laitinen K, Tahtela R, Karkkainen M, Lamberg-Allardt C,et al. Bone mineral density measured by dual-energy X-ray absorptiometry andnovelmarkers of bone formation and resorption in patients on anti-epileptic drugs.J Bone Miner Res 1994;9:631–7.
[10] Lambrinoudaki I, Christodoulakos G, Botsis D. Bisphosphonates. Ann NY Acad Sc2006;1092:397–402.
[11] MauckKF, ClarkeBL. Diagnosis, screening, prevention, and treatment of osteoporosis.Mayo Clin Proc 2006;81(5):662–72.
[12] Gallagher JC, Sai AJ. Molecular biology of bone remodeling: implications for newtherapeutic targets for osteoporosis. Maturitas 2010;65:301–7.
[13] Jarvis CI, Morin AK, Lynch AM. Bisphosphonates for osteoporosis prevention andtreatment. Clin Rev Bone Miner Metab 2005;3(1):3–18.
[14] Tamura T, Aiso K, Johnston KE, Black L, Faught E. Homocysteine, folate, vitamin B-12 and vitamin B-6 in patients receiving antiepileptic drug monotherapy. EpilepsyRes 2000;40:7–15.
[15] Valsamis HA, Arora SK, Labban B, McFarlane SI. Antiepileptic drugs and bonemetabolism. Nutr Metab 2006;3:36–47.
[16] Heath V. Use of antiepileptic drugs implicated in fracture risk. Nat Rev Endocr2010;6:60.
[17] Nissen-Meyer LSH, Svalheim S, Tauboll E, Reppe S, Lekva T, Solberg LB, et al.Levetiracetam, phenytoin, and valproate act differently on rat bonemass, structureand metabolism. Epilepsia 2007;48:1850–60.
[18] Kimball AW, Burnett WT, Doherty DG. Chemical protection against ionizingradiation. I. Sampling methods for screening compounds in radiation protectionstudies with mice. Radiat Res 1957;7:1–12.
[19] Litchfield JT, Wilcoxon F. A simplified method of evaluating dose–effect experi-ments. J Pharmacol Exp Ther 1949;96:99–113.
[20] TenniswoodM, Bird CE, Clark AF. Acid phosphatase: androgen dependent markersof rat prostate. Can J Biochem 1976;54:350–7.
[21] Lowry OH, Rosebrough NJ, Farr AB, Randall RJ. Protein measurement with folin-phenol reagent. J Biol Chem 1951;193:265–75.
[22] Stegemann H, Staldera K. Determination of hydroxyproline. Clin Chim Acta 1967;18(2):267–73.
[23] Koracevic D, Koracevic G, Djordjevic V, Andrejevic S, Cosic V. Method for themeasurement of antioxidant activity in human fluids. J Clin Pathol 2001;54:356–61.
[24] OnoderaK, TakahashiA,WakabayashiH,Kamei J, SakuradaS. Effects ofmenatetrenoneon the bone and serum levels of vitamin K2 (menaquinone derivatives) in osteopeniainduced by phenytoin in growing rats. Nutrition 2003;19:446–50.
[25] Valimaki MJ, Tiihonen M, Laitinen K, Tahtela R, Karkkainen M, Lamberg-Allardt C,et al. Bone mineral density measured by dual-energy X-ray absorptiometry andnovel markers of bone formation and resorption in patients on antiepileptic drugs.Bone Miner Res 1994;9(5):631–7.
[26] Moro-Alvarez MJ, Curiel MD, Piedra C, Marinoso ML, Carrascal MT. Bone diseaseinduced by phenytoin therapy: clinical and experimental study. Eur Neurol 2009;62:219–30.
[27] Kubota F, Kifune A, Shibata N, Akata T, Takeuchi K, Takahashi S, et al. Bone mineraldensity of epileptic patients on long-term antiepileptic drug therapy: aquantitative digital radiography study. Epilepsy Res 1999;33:93–7.
[28] Stephen LJ, McLellan AR, Harrisson JH, Shapiro D, Dominiczak MH, Sills GJ, et al.Bone density and antiepileptic drug: a case-controlled study. Seizure 1999;8:339–42.
[29] Kulak CA, Borba VZ, Silvado CE, Paola L, Seibel MJ, Bilezikian JP, et al. Bone densityand bone turnover markers in patients with epilepsy on chronic antiepileptic drugtherapy. Arq Bras Endocrinol Metabol 2007;51:466–71.
[30] Black DM, Cummings SR, Karpf DB, Cauley JA, Thompson DE, Nevitt MC, et al.Randomised trial of effect of alendronate on risk of fracture in women with
existing vertebral fractures. Fracture Intervention Trial Research Group. Lancet1996;348:1535–41.
[31] Hochberg MC, Ross PD, Black D, Cummings SR, Genant HK, Nevitt MC, et al. Largerincreases in bone mineral density during alendronate therapy are associated witha lower risk of new vertebral fractures in women with postmenopausalosteoporosis. Arthritis Rheum 1999;42:1246–54.
[32] Harris ST,Watts NB, Genant HK, McKeever CD, Hangartner T, Keller M, et al. Effectsof risedronate treatment on vertebral and nonvertebral fractures in women withpostmenopausal osteoporosis: a randomized controlled trial. JAMA 1999;282:1344–52.
[33] Chesnut III CH, Skag A, Christiansen C, Recker R, Stakkestad JA, Hoiseth A, et al.Effects of oral ibandronate administered daily or intermittently on fracture risk inpostmenopausal osteoporosis. J Bone Miner Res 2004;19:1241–9.
[34] Reginster JY, Adami S, Lakatos P, GreenwaldM, Stepan JJ, Silverman SL, et al. Efficacyand tolerability of once-monthly oral ibandronate in postmenopausal osteoporosis:2 year results from the MOBILE study. Ann Rheum Dis 2006;65:654–61.
[35] Silverman SL, Watts NB, Delmas PD, Lange JL, Lindsay R. Effectiveness of bispho-sphonates on nonvertebral and hip fractures in the first year of therapy: therisedronate and alendronate (REAL) cohort study. Osteoporos Int 2007;18:25–34.
[36] Larsen ER, Mosekilde L, Foldspang A. Vitamin D and calcium supplementationprevents osteoporotic fractures in elderly community dwelling residents: a pragmaticpopulation-based 3-year intervention study. J Bone Miner Res 2004;19(3):370–8.
[37] Jackson RD, LaCroix AZ, Gass M,Wallace RB, Robbins J, Lewis CE, et al. Calcium plusvitamin D supplementation and the risk of fractures. N Engl J Med 2006;354(7):669–83.
[38] Yoshiaki D. 5,5-diphenyl-2,4-imidazolidinedion induced osteoclastogenesis incodulture system of osteoblasts and bone marrow cells. Hokkaido J Dental Sci2002;23(1):16–22.
[39] Murakami H, Takahashi N, Sasaki T, Udagawa N, Tanaka S, Nakamura I, et al. Apossible mechanism of the specific action of bisphosphonates on osteoclasts:tiludronate preferentially affects polarized osteoclasts having ruffled borders.Bone 1995;17:137–44.
[40] Jowsey J, Arnaud SB, Hodgson SF, Johnson KA, Beabout JW, Wahner HW. Thefrequency of bone abnormality in patients on anticonvulsant therapy. Electro-encephalogr Clin Neurophysiol 1978;45:341–7.
[41] Vasikaran SD, Glendenning P, Morris HA. The role of biochemical markers of boneturnover in osteoporosis management in clinical practice. Clin Biochem Rev2006;27:119–21.
[42] Christenson RH. Biochemical markers of bone metabolism: an overview. ClinBiochem 1997;30:573–93.
[43] Im G-II, Qureshi SA, Kenney J, Rubash HE, Shanbhag AS. Osteoblast proliferationand maturation by bisphosphonates. Biomater 2004;25:4105–15.
[44] Giuliani N, Girasole G, Pedrazzoni M, Passeri G, Gatti C, Passeri M. Alendronatestimulates b-FGF production and mineralized nodule formation in human osteo-blastic cells and osteoblastogenesis in human bone marrow cultures. J Bone MinerRes 1995;10(1):S171.
[45] Mundy G, Garrett R, Harris S, Chan J, Chen D, Rossini G, et al. Stimulation of boneformation in vitro and in rodents by statins. Science 1999;286:1946–9.
[46] Plotkin LI, Weinstein RS, Parfitt AM, Roberson PK, Manolagas SC, Bellido T.Prevention of osteocyte and osteoblast apoptosis by bisphosphonates andcalcitonin. J Clin Invest 1999;104(10):1363–74.
[47] Delmas PD. Biochemical markers of bone turnover. I: theoretical considerationsand clinical use in osteoporosis. Am J Med 1993;95:11-6S.
[48] KleerekoperM, EdelsonGW. Biochemical studies in the evaluation andmanagementof osteoporosis: current status and future prospects. Endocrin Pract 1996;2:13–9.
[49] Viguet-Carrin S, Garnero P, Delmas PD. The role of collagen in bone strength.Osteoporos Int 2006;17:319–36.
[50] Nishikawa M, Akatsu T, Katayama Y, Yasutomo Y, Kado S, Kugai N, et al.Bisphosphonates act on osteoblastic cells and inhibit osteoclast formation inmouse marrow cultures. Bone 1996;18(1):9–14.
[51] Selhub J. Homocysteine metabolism. Ann Rev Nutr 1999;19:217–46.[52] Tamura T, Aiso K, Johnston KE, Black L, Faught E. Homocysteine, folate, vitamin B-
12 and vitamin B-6 in patients receiving antiepileptic drug monotherapy. EpilepsyRes 2000;40:7–15.
[53] Hoffbrand AV, Necheles TF. Mechanism of folate deficiency in patients receivingphenytoin. Lancet 1968;292(7567):528–30.
[54] Lewis DP, Van Dyke DC, Willhite LA, Stumbo PJ, Berg MJ. Phenytoin–folic acidinteraction. Ann Pharmacother 1995;29(7–8):726–35.
[55] Carl GF, Hudson FZ, McGuire BS. Phenytoin-induced depletion of folate in ratsoriginates in liver and involves a mechanism that does not discriminate folateform. J Nutr 1997;127(11):2231–8.
[56] Carl GF, Smith DB. The effect of chronic phenytoin treatment on tissue folateconcentrations and on the activities of the methyl synthetic enzymes in the rat.J Nutr 1983;113:2368–74.
[57] Kapoor P, Ansari MN, Bhandari U. Modulatory effect of curcumin on methionine-induced hyperlipidemia and hyperhomocysteinemia in albino rats. Ind J Exp Biol2008;46:534–40.
[58] VanMeurs JB, Pluijm SM. Homocysteine level and the risk of osteoporotic fracture.N Engl J Med 2004;350:2033–41.
[59] Elliott JO, Jacobson MP, Haneef Z. Homocysteine and bone loss in epilepsy. Seizure2007;16:22–34.
[60] Loscalzo J. The oxidant stress of hyperhomocyst(e)inemia. J Clin Invest 1996;98:5–7.
[61] Garrett IR, Boyce BF, Oreffo ROC, Bonewald L, Poser J, Mundy GR. Oxygen derivedfree radicals stimulate osteoclastic bone resorption in rodent bone in vitro and invivo. J Clin Invest 1990;85:632–9.
606 S. Khanna et al. / Bone 48 (2011) 597–606
[62] Suda N, Morita I, Kuroda T, Murota S. Participation of oxidative stress in theprocess of osteoclast differentiation. Biochim Biophys Acta 1993;1157:318–23.
[63] Raisanen S, Alatalo S, Ylipahkala H, Halleen J, Cassady A, Hume D, et al. Macrophagesoverexpressing tartrate-resistant acid phosphatase show altered profile of freeradical production and enhanced capacity of bacterial killing. Biochem Biophy ResComm 2005;331:120–6.
[64] LuckeyMM,Gilchrist N, BoneHG, DavieMW, de Villiers TJ,WuM, et al. Therapeuticequivalence of alendronate 35 milligrams once weekly and 5 milligrams daily
in the prevention of postmenopausal osteoporosis. Obstet Gynecol 2003;101:711–21.
[65] Adami S. Bisphosphonate antifracture efficacy. Bone 2007;41:S8–S15.[66] Reginster J, Minne HW, Sorensen OH, Hooper M, Roux C, Brandi ML, et al.
Randomized trial of the effects of risedronate on vertebral fractures in womenwith established postmenopausal osteoporosis. Osteoporos Int 2000;11:83–91.
[67] Drezner MK. Treatment of anticonvulsant drug-induced bone disease. EpilepsyBehav 2004;5:S41–7.
