Discontinuation of carbamazepine due to concerns of long-term consequences of enzyme induction

Discon inua ion o ca bamazepine due o conce ns o long- e m consequences o enzyme induc ion * a Jussi M€ akinen , *Si pa Rainesalo, †‡Jani Rai anen, §Jukka Saa inen, ¶Sa u Sandell, and #Jukka Pel ola Epilepsia Open, 3(3):340–347, 2018 doi: 10.1002/epi4.12227 Jussi M€ akinen, pos doc o al esea che in he Depa men o Neu ology, Uni e si y o Tampe e, neu ologis a Lapland Cen al Hospi al, Finland. SUMMARY Objec i e: T ea men wi h ca bamazepine (CBZ), a po en enzyme induce , is known o a ec he lipid p o ile, s e oid, and i amin D me abolism. Consequen ly, i has been pos ula ed ha pa ien s on CBZ should be swi ched o noninducing an iepilep ic d ugs (AEDs). Howe e , li le is known abou he seizu e ou come ollowing a CBZ swi ch in seizu e- ee pa ien s. We aimed o add ess his issue using a con olled obse a ional s udy design. Me hods: Fi y-eigh pa ien s aking CBZ o ocal epilepsy we e assessed o discon in- uing CBZ ea men due o conce ns o long- e m ad e se-e ec s; 34 discon inued i s he apy and 24 con inued wi h CBZ. Six-mon h seizu e eedom was he p ima y end poin . Fu he mo e, se um samples ( o al choles e ol (TC), low-densi y lipop o ein (LDL), high-densi y lipop o ein (HDL), iglyce ides, sex ho mone–binding globulin (SHBG), ee es os e one, and 25-hyd oxy i amin D le els om be o e and a leas 3 mon hs a e discon inua ion o con inua ion we e ob ained om all pa ien s. Resul s: Seizu e- ee pa ien s had a 5- old ele a ed odds o seizu e ecu ence i CBZ was discon inued (95% con idence in e al [CI 0.51–49.3; p =0.17). A signi ican dec ease in se um le els o TC, LDL, HDL, and SHBG as well as a signi ican inc ease in ha o ee es os e one we e ound in he discon inua ion g oup compa ed wi h hose who con inued CBZ. Nonsigni ican changes in iglyce ides and i amin D le els we e de ec ed. Signi icance: Discon inua ion o CBZ in seizu e- ee pa ien s seems o ca y a mode - a e, bu legi ima e, isk o elapse. Con e sely, ou esul s indica e ha CBZ migh ha e un a o able e ec s on se um le els o TC, LDL, HDL, SHBG, and ee es os- e one. KEY WORDS: An iepilep ic d ugs, T ea men ansi ion, Seizu es. Epilepsy o en equi es long- e m, e en li elong, ea - men wi h an iepilep ic d ugs (AEDs). The a ge o epi- lepsy he apy is always seizu e eedom wi h as ew ad e se e ec s as possible. Ca bamazepine (CBZ) is conside ed as a d ug o choice in ocal epilepsy, 1 al hough a ange o well- ole a ed and e ec i e AEDs ha e become a ailable du ing he pas 2 decades. In addi ion o epilepsy, CBZ is used widely in he ea men o neu algic pain synd omes, mig aine, and some psychia ic indica ions. CBZ is a po en induce o he cy och ome P450 (CYP450) enzyme sys em. 2 In addi ion o hei well-known e ec s on d ug me abolism, CYP450 enzymes a e also in ol ed in endogenous me abolic pa hways, 3 and can in lu- ence bone biochemis y, gonadal s e oids, and choles e ol Accep ed Ap il 23, 2018. *Depa men o Neu ology, Tampe e Uni e si y Hospi al, Tampe e, Finland; †Facul y o Social Sciences, Uni e si y o Tampe e, Tampe e, Finland; ‡UKK Ins i u e o Heal h P omo ion, Tampe e, Finland; §Depa men o Neu ology, Vaasa Cen al Hospi al, Vaasa, Finland; ¶Depa men o Neu ology, Sein€ ajoki Cen al Hospi al, Sein€ ajoki, Finland; and #Depa men o Neu ology, Uni e si y o Tampe e and Tampe e Uni e si y Hospi al, Tampe e, Finland a Add ess co espondence o Jussi M€ akinen, Depa men o Neu ology, Tampe e Uni e si y Hospi al, PO Box 2000, 33521 Tampe e, Finland. E- mail: Jussi.Makinen@ imne . i ©2018 The Au ho s. Epilepsia Open published by Wiley Pe iodicals Inc. on behal o In e na ional League Agains Epilepsy. This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion-NonComme cial-NoDe i s License, which pe mi s use and dis- ibu ion in any medium, p o ided he o iginal wo k is p ope ly ci ed, he use is non-comme cial and no modi ica ions o adap a ions a e made. 340 FULL-LENGTH ORIGINAL RESEARCH syn hesis. 4–7 Second-gene a ion AEDs ha e weak o nonex- is en inducing p ope ies. 8 To da e, a educ ion in o al choles e ol le els has been documen ed a e swi ching om inducing AEDs (CBZ o pheny oin) o lamo igine, le e i ace am, oxca bazepine, opi ama e, o zonisamide. 9– 12 One majo conce n is ha AEDs could con ibu e o he pandemic o ascula disease. 13 Acco dingly, i has been p oposed ha CBZ should no be ega ded as i s -line he - apy in newly diagnosed epilepsy. 14 The impac o hese po en ial dele e ious e ec s o long- e m EI in hose pa ien s cu en ly ecei ing CBZ he apy is unclea , and his cons i u es a signi ican issue in cu en clinical managemen . I me abolic complica ions a e de ec ed, hey migh be ea ed, bu swi ching o a newe AED may be a be e op ion. Such al e a ions a e o en con- side ed and implemen ed in clinical p ac ice wi h he app e- cia ion ha he swi ch also ca ies a isk o elapse. In addi ion, he e a e a numbe o clinical si ua ions equi ing a swi ch om CBZ o a newe AED ( o example, pha ma- co he apy o cance ). Consequen ly, i has been pos ula ed ha pa ien s being ea ed wi h EI AEDs should be swi ched o noninducing AEDs. 15 He e, we p o ide p ac ical in o ma ion ela ed o discon- inua ion o CBZ due o conce ns abou he long- e m e ec s o enzyme induc ion (EI) om 3 di e en iewpoin s: (1) he pa ien 0s pe spec i e, (2) seizu e ou come, and (3) labo a o y pa ame e s (lipids, sex ho mone-binding globulin [SHBG], es os e one, and i amin D). The main objec i e was o cla - i y he impo ance o hese da a on indi idual pa ame e s, when deciding whe he o con inue CBZ ea men . Ma e ials and me hods Pa ien s wi h ocal epilepsy (aged ≥18 yea s) ea ed in Tampe e Uni e si y Hospi al we e iden i ied om he hos- pi al pa ien egis y on 31 Decembe 2014 using In e na- ional Classi ica ion o Diseases (ICD-10) diagnos ic codes o ocal and unclassi iable ocal epilepsy (G40.1X, G40.2X and G40.9). In addi ion, pa ien s om Sein€ ajoki Cen al Hospi al and Vaasa Cen al Hospi al we e iden i- ied. We included pa ien s who we e cu en ly being ea- ed wi h CBZ (mono he apy o poly he apy) and whose ea ing epilep ologis had been conside ing whe he o dis- con inue CBZ due o conce ns abou he long- e m e ec s linked wi h EI. This si ua ion was de ined as he baseline. Those who we e swi ched om CBZ o some o he AED due o unsa is ac o y seizu e con ol, ad e se e en s, d ug in e ac ions, o any eason o he han he possible long- e m consequences o enzyme induc ion, we e excluded. The discon inua ion g oup consis ed o 2 subg oups: (1) con e sion om CBZ o a newe AED o (2) slow wi h- d awal o CBZ wi hou con e sion o any o he AED. Pa ien s who p e e ed o con inue on CBZ we e designa ed as a con ol g oup o compa ison wi h hose unde going CBZ discon inua ion. The decisions conce ning discon inu- a ion we e made pu ely on clinical g ounds. In all cases, an indi idual he apeu ic plan was de ised by he ea ing epilep ologis o he pa ien 0s bene i . The a e o ini ia ion o he apy wi h he new AED and ape ing o om CBZ we e pu sued a he disc e ion o he ea ing epilep ologis and we e indi idualized o each pa ien . The minimum a - ge dose o he new AED was 800 mg daily o eslica - bazepine ace a e, 200 mg daily o lacosamide, 1,000 mg daily o le e i ace am, and 900 mg o gabapen in. We e ospec i ely e iewed pa ien backg ounds, como bidi- ies, cu en and p e ious AED use, and seizu e equency om he pa ien eco ds. Re ac o y epilepsy was quali ied as ha ing pe sis en seizu es a e ials wi h a leas 2 AEDs a maximally ole a ed doses (sequen ially o in combina ion he apy). The use o o he po en -inducing AEDs (pheny oin, p imidone, and phenoba bi al) is ex e- mely limi ed in ou dis ic and he e o e was no included in he cu en s udy. In ou ins i u ions, pa ien s who a e es ablished on EI AEDs a e egula ly sc eened o associa ed long- e m p ob- lems including os eopo osis, hype lipidemia, and sexual dys unc ion. As a pa o his p ocess, we measu e he ol- lowing se um a iables: SHBG, ee es os e one o male pa ien s, 25-hyd oxy i amin D, o al choles e ol, low-den- si y lipop o ein (LDL), high-densi y lipop o ein (HDL), and iglyce ides. Among he women wi h epilepsy, sex ho - mone p o iles a e no ou inely measu ed due o compli- ca ed in e p e a ions wi h espec o he mens ual cycle. In he case o CBZ discon inua ion, con ol blood samples we e ob ained a leas 3 mon hs a e he las dose o CBZ in o de o obse e he possible imp o emen in labo a o y pa ame e s a e de-induc ion. A 3-mon h imespan was designa ed o ensu e su icien ime o comple e de-induc- ion o he CYP450 sys em. In case o CBZ con inua ion, blood samples we e con olled egula ly as a pa o he clin- ical ou ine. The design o he s udy en ailed each pa ien ha ing blood d awn a e a minimum o 12 hou s o as ing Key Poin s •The e has been specula ion abou whe he pa ien s being ea ed wi h inducing AEDs should be swi ched o noninducing AEDs •A signi ican dec ease in se um le els o TC, LDL, HDL, and SHBG and a signi ican inc ease in ee es os e one le els we e ound in he CBZ discon inu- a ion g oup compa ed wi h hose who con inued wi h CBZ •Discon inua ion o CBZ in seizu e- ee pa ien s seems o ca y a mode a e, bu legi ima e isk o elapse •Wi h ega d o he po en ial o ch onic ad e se-e ec s ela ed o EI, he p ac ice o swi ching pa ien s om CBZ o a noninducing AED migh be wo hy o con- side a ion Epilepsia Open, 3(3):340–347, 2018 doi: 10.1002/epi4.12227 341 Discon inua ion o Ca bamazepine on 2 occasions: i s a baseline, while aking CBZ; a second ime a 3 mon hs a e he las dose o CBZ. Gi en he d ug in e ac ions be ween CBZ and s a ins, he lipid analyses we e pe o med wi h exclusion o he s a in use s. Seizu e ou come was es ablished o 12 mon hs be o e and o he 6 mon hs a e baseline. We de ined seizu e ou - come as ei he (1) ha ing no seizu es o 6 mon hs, which signi ied ha he pa ien was seizu e- ee; o (2) ha ing expe ienced a seizu e on a he apeu ic dose o hei AED du ing he 6-mon h pe iod, which indica ed he pa ien was no seizu e- ee. Isola ed au as we e no conside ed in his assessmen . We also excluded seizu es occu ing du ing he i a ion pe iod o he new AED. Desc ip i e s a is ics (means and anges [min o max] o equencies and p opo ions) we e u ilized o summa ize pa ien s’cha ac e is ics a baseline. Changes in se um sam- ples we e calcula ed as he di e ence be ween he end o he ollow-up and baseline and we e epo ed as means and s anda d de ia ions. We combined 5 se um samples (HDL, LDL, iglyce ides, SHBG, and i amin D) o e alua e hei o e all signi icance in pa ien s. To combine alues om hese 5 di e en scaled a iables, we i s s anda dized each o hem, ha is, escaled hem o ha e a mean o ze o and a s anda d de ia ion one. The alue o he s anda dized a i- able (also called z-sco e) ep esen s i s dis ance om he mean in s anda d de ia ion uni s. Fo example, a s anda d- ized alue o 0.5 indica es ha he alue is hal a s anda d de ia ion below he mean. Thus, a posi i e alue o he sum o 5 s anda dized a iables (z-sum) indica es ha he alue is abo e he mean o he z-sum. Be o e summa ion, we changed he sign o a iables LDL, iglyce ides, and SHBG in o de o code all 5 a iables in same o de , ha is, la ge posi i e alue indica es a “be e si ua ion” o all a iables. Chi-squa e es was used o analyze di e ences be ween g oups and ca ego ical a iables, his being based on he assump ion o no mali y; S uden ’s - es o Mann-Whi ney U- es was used o con inuous a iables. Bina y logis ic eg ession analysis was used o es ima e he associa ion be ween g oups and seizu e ecu ence a 6 mon hs a e baseline. No essen ial changes in he es ima e o g oup a i- able we e ound a e adjus men s o age, gende , and he numbe o p io AEDs and he e o e only esul s om unad- jus ed models we e epo ed. A e i ing logis ic eg ession models, p edic ed p obabili ies o seizu e ecu ence 6 mon hs a e baseline we e calcula ed by con e ing o odds. All analyses we e conduc ed using S a a s a is ical so - wa e e sion 13.1 (S a aCo p, College S a ion, Texas, U.S.A.). Fo all s a is ical es s, p- alue <0.05 was consid- e ed s a is ically signi ican . This was an obse a ional, nonin e en ional e ospec- i e s udy, which does no equi e e hics commi ee app o al acco ding o Finnish Law on Resea ch. Access o pa ien eco ds was based on a decision made by he Head o Science Cen e, Tampe e Uni e si y Hospi al Resea ch and Inno a ion Se ices, Science Cen e . Resul s O he 58 pa ien s pa icipa ing in he s udy, 24 decided o con inue wi h CBZ, 10 we e wi hd awn (7 o hem seizu e- ee) om CBZ wi hou swi ching o some o he AED, and 24 (13 o hem seizu e- ee) we e con e ed om CBZ o some o he AED (8 o eslica bazepine ace a e, 8 o lacosa- mide, 7 o le e i ace am, and one o gabapen in). The main easons a ec ing indi idual decisions o con inue wi h CBZ we e as ollows; 11 we e anxious ha he e would be sei- zu e- elapse, 7 we e a aid o losing hei job and/o d i e 0s license should a seizu e occu , 2 we e no conce ned abou he long- e m e ec s o EI, wi h he inal 4 subjec s deciding o easons bes o known hemsel es. The baseline da a o all pa ien s appea in Table 1. The e we e signi ican di e ences be ween he g oups wi h espec o seizu e eedom and mean ee es os e one le els. Recu en seizu es we e mo e equen , and he es os e one le el was lowe in he discon inua ion g oup a baseline compa ed o con inua ion g oup. The ime be ween he i s and second blood samplings in hese pa ien s anged om 91 o 334 days (mean 141 days). Compa ed o hose who con inued he CBZ ea men , pa ien s in he CBZ discon inua ion g oup exhibi ed signi i- can dec eases in se um o al choles e ol (16%), HDL (11%), LDL (18%), and SHBG (18%) concen a ions (Table 2.) In men, he ee es os e one le el was signi i- can ly inc eased (39%) in he CBZ discon inua ion g oup compa ed o hose who con inued wi h CBZ medica ion. The e we e no signi ican changes in he se um concen a- ions o iglyce ide and i amin D. Eslica bazepine ace a e migh i sel ha e some impac on lipids, o pe haps o he se ologic ma ke s. The e o e, we e-analyzed he da a ega ding all he se ologic ma k- e s a e exclusion o hose who we e swi ched o eslica - bazepine ace a e. The esul s emained unchanged a e exclusion (da a no shown). Two pa ien s had a spo adic seizu e du ing he i a ion pe iod o he new d ug. In con- as , none had wi hd awal seizu es due o ape ing o he CBZ medica ion. Da a om hese seizu es a e no included he e. Table 3 demons a es unadjus ed odds a ios o he a ious subg oups ela i e o hei seizu e s a us a baseline. The logis ic eg ession model was adjus ed o co a ia es (age, gende , and numbe o p io AEDs), bu his did no p oduce esul s ha di e ed ma k- edly om hose ob ained wi h he unadjus ed model (da a no shown). Table 4 p esen s p obabili ies o seizu e ecu ence 6 mon hs a e baseline in he a ious sub- g oups o p e iously seizu e- ee pa ien s. P obabili ies a e based on calcula ions made om he esul s o he logis ic eg ession model (Table 3). Epilepsia Open, 3(3):340–347, 2018 doi: 10.1002/epi4.12227 342 J. M€ akinen e al. Discussion The majo ea u e o cu en s udy is he sys ema ic assessmen on he seizu e ou come o CBZ discon inua ion in a ious subg oups. In addi ion, we show he e ha CBZ discon inua ion p oduces a b oad spec um o signi ican changes in se um concen a ions o o al choles e ol, HDL, LDL, SHBG, and es os e one, whe eas p e ious s udies ha e ocused on single me abolic pa hways. Despi e he el- a i ely modes sample size, he signi icance o he indings a es s o he obus na u e o he e ec . The o e all conse- quence o hese changes would be expec ed o esul in a conside able decline in he isk o ischemic ascula dis- ease and sexual dys unc ion in men. In he cu en s udy, app oxima ely 40% o he pa ien s we e ul ima ely no willing o discon inue CBZ e en hough hey had ini ially exp essed acquiescence owa d an e alua- ion o me abolic side-e ec s ela ed o CBZ. Mos pa ien s had been seizu e- ee wi h CBZ o decades wi hou expe i- encing any no iceable side e ec s. On he o he hand, aging is he mos impo an isk ac o o ascula e en s. How- e e , a signi ican p opo ion o pa ien s decided o con inue CBZ despi e hei app ecia ion o he possible isks ela ed o he long- e m e ec s o EI. The impo ance o good com- munica ion be ween pa ien and ea ing physician is unde - sco ed in hese si ua ions. The e we e some di e ences in he baseline cha ac e is- ics o he s udy pa ien s ela i e o CBZ s a us a baseline (con inua ion s discon inua ion). The p opo ion o sei- zu e- ee pa ien s was signi ican ly highe in hose who con- inued CBZ and who we e anxious o he possibili y o he e-appea ance o seizu es in hese cu en ly seizu e- ee subjec s. The pe cen age o ee es os e one was lowe in he CBZ discon inua ion g oup, sugges ing ha especially men wi h low es os e one le els migh be mo e willing o end CBZ medica ion in compa ison wi h hei coun e pa s wi h no mal es os e one le els. The inclusion o a con ol g oup in ou s udy has added an impo an me hodologic ea u e missing om all p e ious ou come in es iga ions, wi h 2 excep ions done by he same g oup. 16,17 Howe e , in he s udy by Wang e al., 16 all swi ched pa ien s we e ini ially aking pheny oin o CBZ, whe eas he con ols we e being ea ed wi h a di e se g oup o 10 di e en AEDs. We compa ed he pa ien s who dis- con inued CBZ o hose aking CBZ who con inued aking CBZ. One o he majo ea u es o he cu en s udy is o p ac i- cal signi icance because we we e able o calcula e p obabili- ies o seizu e ecu ence among a ious subg oups (Table 4). The ecu ence a e a e CBZ wi hd awal esul ed in an app oxima ely 24 pe cen age poin inc emen- al addi ional isk o seizu e ecu ence. Simila ly, con e - ing CBZ o ano he AED in seizu e- ee pa ien s esul ed in an 11 pe cen age poin addi ional isk o ecu en seizu es, which is in line wi h ecen s udy. 17 We also calcula ed he odds o seizu e ecu ence in seizu e- ee pa ien s, bu he absolu e di e ences migh be mo e in e es ing om he clinician’s pe spec i e. Mos o he epidemiological da a demons a e ha pa ien s wi h epilepsy ha e an inc eased isk o de eloping ca dio ascula and ce eb o ascula disease compa ed o he gene al popula ion. 18–22 In addi ion, ca o id media in ima hickness is signi ican ly inc eased in pa ien s wi h epilepsy, pa icula ly among hose aking CBZ. 23 Fu he mo e, Sil- lanp€ a€ a e al. 24 conduc ed a popula ion-based coho s udy and demons a ed ha he e was a s iking inc ease in mag- ne ic esonance imaging (MRI) abno mali ies ela ed o ce eb o ascula disease in pa ien s wi h epilepsy compa ed o heal hy con ols a he age o 45 yea s. Following an Table 1. Baseline cha ac e is ics o he s udy pa ien s CBZ s a us a baseline p- alueCon inua ion Discon inua ion N2434 Age in yea s ( ange) 52.6 (25–69) 49.1 (29–78) 0.29 a Female (%) 13 (54.2) 19 (55.9) 0.90 b S a in use (%) 2 (8.3) 5 (14.7) 0.46 b Numbe o p io AEDs (%) 0.11 b 0 10 (41.7) 13 (38.2) 1–2 9 (37.5) 6 (17.7) 3+5 (20.8) 15 (44.1) Re ac o y (%) 9 (37.5) 15 (44.1) 0.61 b Numbe o concomi an AEDs (%) 0.62 b 0 12 (50.0) 15 (44.1) 1 7 (29.2) 14 (41.2) 2–3 5 (20.8) 5 (14.7) Du a ion o epilepsy, yea s ( ange) 34.4 (2–53) 30.3 (1–62) 0.32 a Du a ion o CBZ, yea s ( ange) 29.0 (2–48) 23.7 (1–49) 0.14 a Daily dose o CBZ, mg ( ange) 810 (400–1500) 750 (400–1600) 0.47 a Seizu e ee (%) 21 (87.5) 20 (58.8) 0.018 b To al choles e ol, mM(SD) 5.7 (1.2) 5.9 (1.3) 0.59 a HDL, mM(SD) 1.8 (0.6) 2.0 (0.6) 0.25 a LDL, mM(SD) 3.6 (1.0) 3.7 (1.0) 0.72 a T iglyce ide, mM(SD) 1.3 (0.8) 1.1 (0.6) 0.65 c SHBG, nM(SD) 115.5 (87.1) 100.4 (48.2) 0.97 c F ee es os e one, pM(SD) d 212.8 (73.8) 156.9 (57.5) 0.046 c Vi amin D, nM(SD) 80.2 (34.4) 78.1 (35.7) 0.83 a z-sco e sum e 0.36 (2.66) 0.25 (2.01) 0.32 a AEDs, an iepilep ic d ugs; CBZ, ca bamazepine; HDL, high-densi y lipop o- ein; LDL, low-densi y lipop o ein; SD, s anda d de ia ion, SHBG, sex ho mone –binding globulin. a S uden ’s - es . b Chi-squa e es . c Mann-Whi ney U- es . d n=26 (men only). e The sum o 5 a iables (z-sum) was calcula ed so ha he g ea e alue o z- sum co esponds o be e o al si ua ion. Epilepsia Open, 3(3):340–347, 2018 doi: 10.1002/epi4.12227 343 Discon inua ion o Ca bamazepine ex ensi e me a-analysis, Lossius e al. 25 concluded ha a dec ease in LDL o 0.51 mM, simila o he decline de ec ed in ou pa ien s, could educe o e all mo ali y by 5 pe cen - age poin s and ca dio ascula mo ali y by 11%. 26 Based on he abo e indings, a he oscle o ic ascula disease appea s o be a genuine isk in he epilep ic popula ion. Howe e , con adic o y da a do exis . 27–30 We obse ed ha he e was a dec ease in se um con- cen a ions o o al choles e ol, HDL, and LDL ollowing discon inua ion o CBZ. Fu he mo e, he e was a small, s a is ically insigni ican , decline in iglyce ide le els seen a e CBZ discon inua ion. These esul s a e a he simila o hose seen when pa ien s we e ei he swi ched om CBZ o some o he AED o CBZ was wi h- d awn. 9,10,18,25 The declines in HDL a e likely o be mo e han compensa ed by he nega i e e ec s on p o- a he ogenic ma ke s. 11 Some a iabili y was seen in he CBZ con inua ion g oup, p esumably e lec ing he inhe - en luc ua ions in hese pa ame e s. Mos s a ins a e ex ensi ely me abolized by he CYP sys- em and he e would be expec ed o educe se um le els o hese d ugs in he p esence o an enzyme induce . 15,18 How- e e , pa ien s aking s a ins had been excluded om all p e- ious s udies in es iga ing his opic. In he cu en s udy, some o he pa ien s ecei ing s a ins displayed mode a e declines in he lipid p o ile, o he s exhibi ed a he majo declines, o example, 4.5 mM o al choles e ol o 3.7 mM LDL (Fig. 1). Al hough he numbe o pa ien s wi h s a ins was low and cau ion is necessa y when conside ing he clin- ical implica ions o his inding, one migh specula e ha pa ien s ecei ing s a in he apy should a oid CBZ ea men . Table 3. Odds o seizu e ecu ence 6 mon hs a e baseline, among a ious subg oups ela i e o seizu e s a us a baseline (seizu e- ee o no seizu e- ee) Compa ison Odds a io 95% CI p- alue Seizu e- ee pa ien s a baseline, CBZ discon inue e sus con inue (n =41) 5.00 0.51–49.3 0.17 Seizu e- ee pa ien s a baseline, CBZ wi hd awal e sus con inue (n =28) 8.00 0.60–106.9 0.12 Seizu e- ee pa ien s a baseline, CBZ swi ch o o he AED e sus con inue (n =34) 3.64 0.30–44.8 0.31 AED, an iepilep ic d ug; CBZ, ca bamazepine; CI, con idence in e al. Table 4. P obabili ies o seizu e ecu ence 6 mon hs a e baseline among a ious subg oups ela i e o seizu e s a us a baseline (seizu e- ee o no -seizu e- ee) Compa ison Seizu e ecu ence p obabili ies Di e ence Seizu e- ee pa ien s a baseline, CBZ discon inue e sus con inue (n =41) 20.0% e sus 4.8% 15.2 PP Seizu e- ee pa ien s a baseline, CBZ wi hd awal e sus con inue (n =28) 28.6% e sus 4.8% 23.8 PP Seizu e- ee pa ien s a baseline, CBZ swi ch o o he AED e sus con inue (n =34) 15.4% e sus 4.8% 10.6 PP AED, an iepilep ic d ug; CBZ, ca bamazepine; PP, pe cen age poin . Table 2. Labo a o y da a a e baseline (sampling 2) and compa ison o labo a o y pa ame e s be ween sampling one and sampling wo. S a in use s we e excluded om he lipid analyses CBZ s a us a baseline Change om sampling 1 o sampling 2 Con inue (n =24) Discon inue (n =34) p- alue Con inue (n =24) Discon inue (n =34) p- alue TC, mM(SD) c 5.8 (1.2) 5.1 (1.0) 0.033 a 0.05 (0.51) 0.84 (0.78) <0.001 a HDL, mM(SD) c 1.9 (0.7) 1.8 (0.5) 0.70 a 0.02 (0.16) 0.21 (0.34) 0.004 a LDL, mM(SD) c 3.5 (1.1) 3.1 (0.9) 0.18 a 0.05 (0.64) 0.64 (0.90) 0.006 a T iglyce ide, mM(SD) c 1.18 (0.75) 0.94 (0.40) 0.41 b 0.06 (0.32) 0.14 (0.42) 0.13 b SHBG, nM(SD) 124.7 (88.8) 82.1 (40.3) 0.042 b 9.1 (18.0) 18.4 (39.9) <0.001 b FT, pM(SD) d 212.9 (80.2) 218.1 (93.6) 0.92 b 0.09 (22.5) 61.2 (91.8) 0.017 b Vi amin D, nM(SD) 80.3 (31.3) 81.4 (27.2) 0.89 a 0.1 (19.2) 3.3 (22.4) 0.58 a z-sco e sum e 0.84 (2.84) 0.59 (2.06) 0.030 a 0.48 (1.65) 0.34 (1.61) 0.064 a FT, ee es os e one; HDL, high-densi y lipop o ein; LDL, low-densi y lipop o ein; SD, s anda d de ia ion; SHBG, sex ho mone–binding globulin; TC, o al choles e ol. a S uden ’s - es . b Mann-Whi ney U- es . c n=51 (s a in use s excluded). d n=26 (men only). e The sum o 5 a iables (z-sum) was calcula ed so ha he g ea e alue o z-sum co esponds o be e o al si ua ion. Epilepsia Open, 3(3):340–347, 2018 doi: 10.1002/epi4.12227 344 J. M€ akinen e al. Figu e 1. Change in labo a o y pa ame e s in pa ien s who con inued wi h CBZ and in hose wi h CBZ discon inua ion. Each ba shows he abso- lu e change in he labo a o y pa ame e be ween he i s and second samplings o indi idual subjec s, wi h 24 pa ien s who con inued CBZ shown in whi e on he le , and he 34 pa ien s who discon inued CBZ in g ay on he igh . A black ba indica es a s a in use . Epilepsia Open ILAE Epilepsia Open, 3(3):340–347, 2018 doi: 10.1002/epi4.12227 345 Discon inua ion o Ca bamazepine We de ec ed a signi ican dec ease in SHBG le els in bo h gende s ollowing CBZ discon inua ion. In addi ion, he e was a signi ican inc ease in he ee es os e one le el in men a e CBZ discon inua ion. This co esponds well wi h clinical expe ience ha some men enjoy imp o emen s in sexual unc ion a e CBZ discon inua ion. Un o u- na ely, s anda dized sexual unc ion ques ionnai es a e no ou inely used in ou ins i u ions, so i was no possible o in es iga e whe he he imp o emen s in se um concen a- ions o SHBG and ee es os e one co ela ed wi h eco - e y om sexual dys unc ion. Howe e , biochemical abno mali ies a e clea ly e iden in men, and hese migh ha e a eal impac on pa ien well-being. The design o ou s udy mean ha we we e no able o assess he sex ho mone p o iles o emales as i is no con enien in clinical p ac ice o eques he pa ien o come o he clinic o p o ide a blood sample a a ixed ime poin ela i e o he mens ual cycle. Unexpec edly, discon inua ion o CBZ, howe e , was no associa ed wi h any inc ease in se um i amin D concen a- ions. Se e al ac o s migh be specula ed o accoun o his phenomenon. Fi s , ex ensi e a iabili y was also seen in he CBZ con inua ion g oup in i amin D le els, which migh be ela ed o seasonal changes o sun exposu e wi h espec o he ime o yea . Second, i amin D supplemen s a e p esc ip ion- ee and widely used in Finland. In he no he n hemisphe e a la i udes g ea e han a ound 40°N (no h o Ba celona), sunligh is no s ong enough o igge he syn hesis o i amin D in he skin om Oc obe o Ma ch. Anecdo ally, i amin D le els seemed o be ele a ed pa icula ly in hose pa ien s wi h s a in co-medica ion in he CBZ discon inua ion g oup (Fig. 1), which migh indi- ca e ha s a ins may inc ease i amin D concen a ions. 31 As a as we a e awa e, no p e ious s udy has epo ed such a comp ehensi e labo a o y panel ela ed o enzyme- inducing (o EI) AEDs. This pa e n o labo a o y es s was ound use ul in helping he clinician o es ima e he o e all e ec s o EI in he indi idual pa ien . Fu he mo e, he nume ic pa ame e s may highligh he long- e m conse- quences o EI o he pa ien in a mo e conc e e manne and assis he pa ien o decide o he /himsel whe he o con- inue wi h CBZ he apy. Some poin s mus be kep in mind when d awing conclu- sions om ou s udy. This was a e ospec i e s udy, no a p ospec i e andomized con olled ial, complica ing he compa abili y o pa ien s in he di e en g oups. Howe e , his clinical ques ion could no be adequa ely answe ed in a double-blind andomized s udy. One po en ial asce ain- men bias migh ha e a ec ed he esul s: hose who swi ched may ha e al eady had some no ion ha hey had a p oblem (e.g., wi h bones o choles e ol), o a amily his o y o p oblems, o make hem conce ned. Mo eo e , because he seizu e da a was ga he ed only om he p e ious yea , we we e unable o elimina e he possibili y o emi ing- elapsing pa e n, which is belie ed o be p esen in up o 16% o pa ien s wi h epilepsy, ha is, he pa ien s luc ua e be ween pe iods o seizu e eedom and ecu ence. 32 Sam- ple size migh be ep esen ed as a limi a ion o seizu e ou - comes, bu ou esul s a e simila o he p io 2 s udies, 16,17 sugges ing ha his is no uly an issue. Fu he mo e, we we e unable o s udy pa ien s s a ing o ecei e CBZ he - apy because his is no compa ible wi h ou daily clinical p ac ice. Finally, we did no accoun o o he ac o s ha migh in luence he changes in labo a o y pa ame e s, such as body weigh , die , exe cise, o smoking habi s. In conclusion, we belie e ha he clinical implica ions o he cu en s udy a e conside able o he gene al heal h o pa ien s wi h epilepsy. Wi h ega d o he po en ial o ch onic ad e se e ec s ela ed o EI, he p ac ice o swi ch- ing CBZ pa ien s o noninducing AED migh be wo h con- side a ion. CBZ is esponsible o an ele a ion in he lipid p o ile, al e a ions in male ep oduc i e unc ion, and po en ial d ug in e ac ions (especially wi h s a ins). The e- o e, he use o CBZ is p oblema ic, pa icula ly in pos - s oke epilepsy and in pa ien s wi h an inc eased isk o ascula disease. On he o he hand, one canno s a e ha all CBZ should be swi ched o ano he non-EI AED; o exam- ple, i he pa ien has achie ed seizu e- eedom bu is subse- quen ly ound o ha e side e ec s on impo an me abolic pa hways, he e is none heless a eal isk o seizu e ecu - ence i he/she should be swi ched o some o he AED. All o hese challenges migh be a oided by assigning he app op ia e d ug in he i s place wi h espec o he la es expe opinion on ea men o epilepsy, which compa ed o ea lie su eys, highligh s a mo e away om CBZ as he d ug o choice. 33 Acknowledgmen s All au ho s mee he In e na ional Commi ee o Medical Jou nals Edi- o s (ICMJE) c i e ia o au ho ship and ha e gi en inal app o al o he manusc ip o be published. The au ho s con i m ha hey ha e ead he Jou nal0s posi ion on issues in ol ed in e hical publica ion and a i m ha his epo is consis en wi h hose guidelines. Con lic o in e es and sou ces o unding Jussi M€ akinen has ecei ed suppo o a el cong esses om Biogen- Idec, Boeh inge -Ingelheim, and Eisai; ecei ed speake hono a ia om Boeh inge -Ingelheim; ecei ed esea ch unding om Finnish B ain Foun- da ion s , Finnish Cul u al Founda ion, Pi kanmaa Regional Fund, Finnish Epilepsy Associa ion, Finnish No wegian Medical Founda ion, and O ion Resea ch Founda ion s ; and pa icipa ed in an ad iso y boa d o Eisai. Si pa Rainesalo has ecei ed speake hono a ia om FennoMedical, O ion Pha ma, UCB, and ecei ed suppo o a el o cong esses om Abb ie and UCB. Jukka Saa inen has pa icipa ed in scien i ic ad iso y boa ds o As a- Zeneca, Baye , Biogen-Idec, Sano i-Genzyme, and Shi e; has ecei ed unding o a el om Abb ie, Baye , Biogen-Idec, Sano i-Genzyme, Med onic, Me ck, Roche, Shi e, and Te a; has ecei ed speake hono a ia om Biogen-Idec, Boeh inge Ingelheim, No a is, Sano i-Genzyme, O ion, and Shi e; and has ecei ed esea ch suppo om Sano i-Genzyme and Baye . Jukka Pel ola has pa icipa ed in clinical ials o Eisai, UCB, and Bial; ecei ed esea ch g an s om Eisai, Med onic, UCB, and Cybe onics; Epilepsia Open, 3(3):340–347, 2018 doi: 10.1002/epi4.12227 346 J. M€ akinen e al. ecei ed speake hono a ia om Cybe onics, Eisai, Med onic, O ion Pha ma, and UCB; ecei ed suppo o a el cong esses om Cybe onics, Eisai, Med onic, and UCB; and pa icipa ed in ad iso y boa ds o Cybe onics, Eisai, Med onic, UCB, and P ize . The emaining au ho s ha e no con lic s o in e es . Re e ences 1. Ka ceski S, Mo ell MJ, Ca pen e D. T ea men o epilepsy in adul s: expe opinion. Epilepsy Beha 2005;7:S1–S67. 2. Pa salos PN, Duncan JS, Sho on SD. E ec o he emo al o indi id- ual an iepilep ic d ugs on an ipy ine kine ics, in pa ien s aking poly- he apy. B J Clin Pha macol 1988;26:253–259. 3. Nebe DW, Russell DW. Clinical impo ance o he cy och omes P450. Lance 2002;360:1155–1162. 4. He zog AG, D islane FW, Schome DL, e al. Di e en ial e ec s o an iepilep ic d ugs on sexual unc ion and ho mones in men wi h epi- lepsy. Neu ology 2005;65:1016–1020. 5. Pack AM, Mo ell MJ, Ma cus R, e al. Bone mass and u no e in women wi h epilepsy on an iepilep ic d ug mono he apy. Ann Neu ol 2005;57:252–257. 6. Min ze S, Boppana P, Togu i J, e al. Vi amin D le els and bone u n- o e in epilepsy pa ien s aking ca bamazepine o oxca bazepine. Epilepsia 2006;47:510–515. 7. Nikolaos T, S ylianos G, Ch yssoula N, e al. The e ec o long- e m an iepilep ic ea men on se um choles e ol (TC, HDL, LDL) and iglyce ide le els in adul epilep ic pa ien s on mono he apy. Med Sci Moni 2004;10:MT50–MT52. 8. Ande son GD. Pha macogene ics and enzyme induc ion/inhibi ion p ope ies o an iepilep ic d ugs. Neu ology 2004;63:S3–S8. 9. Isoj€ a i J, Paka inen AJ, Lukka inen O, e al. Li e enzyme induc ion and se um lipid le els a e eplacemen o ca bamazepine wi h oxca - bazepine. Epilepsia 1994;35:1217–1220. 10. Min ze S, Skidmo e CT, Abidin CJ, e al. E ec s o an iepilep ic d ugs on lipids, homocys eine, and C- eac i e p o ein. Ann Neu ol 2009;65:448–456. 11. Min ze S, Skidmo e CT, Rankin SJ, e al. Con e sion om enzyme- inducing an iepilep ic d ugs o opi ama e: e ec s on lipids and c- eac- i e p o ein. Epilepsy Res 2012;98:88–93. 12. Min ze S, Mille R, Shah K, e al. Long- e m e ec o an iepilep ic d ug swi ch on se um lipids and C- eac i e p o ein. Epilepsy Beha 2016;58:127–132. 13. Beaglehole R, Boni a R, Ho on R, e al. P io i y ac ions o he non- communicable disease c isis. Lance 2011;377:1438–1447. 14. Min ze S, Ma son RT. Should enzyme-inducing an iepilep ic d ugs be conside ed i s -line agen s? Epilepsia 2009;50:S42–S50. 15. B odie MJ, Min ze S, Pack AM, e al. Enzyme induc ion wi h an iepilep ic d ugs: cause o conce n? Epilepsia 2013;54:11–27. 16. Wang SP, Min ze S, Skidmo e CT, e al. Seizu e ecu ence and emission a e swi ching an iepilep ic d ugs. Epilepsia 2013;54:187– 193. 17. Finamo e JM, Spe ling MR, Zhan T, e al. Seizu e ou come a e swi ching an iepilep ic d ugs: a ma ched, p ospec i e s udy. Epilepsia 2016;57:1294–1300. 18. Cand illi SD, Manjuna h R, Da is KL, e al. The associa ion be ween an iepilep ic d ug and HMG-CoA educ ase inhibi o co-medica ion and choles e ol managemen in pa ien s wi h epilepsy. Epilepsy Res 2010;91:260–266. 19. Annege s JF, Hause WA, Shi s SB. Hea disease mo ali y and mo - bidi y in pa ien s wi h epilepsy. Epilepsia 1984;25:699–704. 20. Nilsson L, Tomson T, Fa ahmand BY, e al. Cause-speci ic mo ali y in epilepsy: a coho s udy o mo e han 9,000 pa ien s once hospi al- ized o epilepsy. Epilepsia 1997;38:1062–1068. 21. Gai a zis A, Ca oll K, Majeed A, e al. The epidemiology o he como bidi y o epilepsy in he gene al popula ion. Epilepsia 2004;45:1613–1622. 22. Jansky I, Hallq is J, Tomson T, e al. Inc eased isk and wo se p og- nosis o myoca dial in a c ion in pa ien s wi h p io hospi aliza ion o epilepsy –The S ockholm Hea Epidemiology P og am. B ain 2009;132:2798–2804. 23. Hamed SA, Hamed EA, Hamdy R, e al. Vascula isk ac o s and oxida i e s ess as independen p edic o s o asymp oma ic a he oscle- osis in adul pa ien s wi h epilepsy. Epilepsy Res 2007;74:183–192. 24. Sillanp€ a€ a M, An inen A, Rinne JO, e al. Childhood-onse epilepsy i e decades la e . A p ospec i e popula ion-based coho s udy. Epilepsia 2015;56:1774–1783. 25. Lossius MI, Nakken KO, Mowinckel P, e al. Fa o able change o lipid p o ile a e ca bamazepine wi hd awal. Ac a Neu ol Scand 2016;134:219–223. 26. Choles e ol T ea men T ialis s0(CTT) Collabo a ion. Baigen C, Blackwell L, Embe son J, e al. E icacy and sa e y o mo e in ensi e lowe ing o LDL choles e ol: a me a-analysis o da a om 170,000 pa icipan s in 26 andomized ials. Lance 2010;376:1670–1681. 27. Luoma PV, So aniemi EA, A an o AJ. Se um LDL choles e ol, he LDL/HDL choles e ol a io and li e mic osomal enzyme induc ion e alua ed by an ipy ine kine ics. Scand J Clin Lab In es 1983;43:671–675. 28. Luoma PV, So aniemi EA, Pelkonen RO, e al. Se um low-densi y lipop o ein and high-densi y lipop o ein choles e ol, and li e size in subjec s on d ugs inducing hepa ic mic osomal enzymes. Eu J Clin Pha macol 1985;28:615–618. 29. Sudhop T, Baue J, Elge CE, e al. Inc eased high-densi y lipop o ein choles e ol in pa ien s wi h epilepsy ea ed wi h ca bamazepine: a gende - ela ed s udy. Epilepsia 1999;40:480–484. 30. Nakken KO, Ko ns ad S. Do males 30-50 yea s o age wi h ch onic epilepsy and on long- e m an icon ulsan medica ion ha e lowe - han- expec ed isk o de eloping co ona y hea disease? Epilepsia 1998;39:326–330. 31. Mazidi M, Rezaie P, Va anpa as H, e al. E ec o s a ins on se um i amin D concen a ions: a sys ema ic e iew and me a-analysis. Eu J Clin In es 2017;47:93–101. 32. B odie MJ, Ba y SJE, Bamagous GA, e al. Pa e ns o ea men esponse in newly diagnosed epilepsy. Neu ology 2012;78:1548–1554. 33. Shih JJ, Whi lock JB, Chima o N, e al. Epilepsy ea men in adul s and adolescen s: expe opinion, 2016. Epilepsy Beha 2017;69:186– 222. Epilepsia Open, 3(3):340–347, 2018 doi: 10.1002/epi4.12227 347 Discon inua ion o Ca bamazepine