ARTICLE OPEN ACCESS

Genetic variation in CFH predictsphenytoin-induced maculopapular exanthemain European-descent patientsMarkMcCormack PhDHongshengGui PhDAndres Ingason PhDDougSpeed PhDGalen EBWright PhD

Eunice J Zhang PhD Rodrigo Secolin PhD Clarissa YasudaMD PhDMaxwell KwokMPhil StefanWolkingMD

Felicitas Becker Sarah Rau MPH Andreja Avbersek MD Kristin Heggeli BSc Costin Leu PhD

Chantal Depondt PhDGraeme J Sills PhD AnthonyGMarsonMD FRCP Pauls AuceMDMartin J BrodieMD

FRCP Ben Francis PhDMichael R Johnson DPhil FRCP Bobby PC Koeleman PhD Pasquale StrianoMD PhD

Antonietta Coppola MD PhD Federico Zara PhD Wolfram S Kunz PhD Josemir W Sander FRCP

Holger Lerche MD Karl Martin Klein MD PhD Sarah Weckhuysen MD PhD Martin Krenn MD

Larus J Gudmundsson BSc Kari Stefansson PhD Roland Krause PhD Neil Shear MD Colin JD Ross PhD

and Norman Delanty FRCPI for the EPIGEN Consortium Munir Pirmohamed FRCP

and Bruce C Carleton PharmD for the Canadian Pharmacogenomics Network for Drug Safety

Fernando Cendes MD PhD Iscia Lopes-Cendes MD PhD Wei-ping Liao MD PhD Terence J OrsquoBrien PhD

and Sanjay M Sisodiya FRCP for the EpiPGX Consortium Stacey Cherny PhD Patrick Kwan PhD

and Larry Baum PhD for the International League Against Epilepsy Consortium on Complex Epilepsies

and Gianpiero L Cavalleri PhD

Neurologyreg 201890e332-e341 doi101212WNL0000000000004853

Correspondence

Prof Cavalleri

gcavallerircsiie

AbstractObjectiveTo characterize among European and Han Chinese populations the genetic predictors ofmaculopapular exanthema (MPE) a cutaneous adverse drug reaction common to antiepilepticdrugs

MethodsWe conducted a case-control genome-wide association study of autosomal genotypes in-cluding Class I and II human leukocyte antigen (HLA) alleles in 323 cases and 1321 drug-tolerant controls from epilepsy cohorts of northern European and Han Chinese descentResults from each cohort were meta-analyzed

ResultsWe report an association between a rare variant in the complement factor Hndashrelated 4(CFHR4) gene and phenytoin-induced MPE in Europeans (p = 45 times 10ndash11 odds ratio [95confidence interval] 7 [32ndash16]) This variant is in complete linkage disequilibrium witha missense variant (N1050Y) in the complement factor H (CFH) gene In addition our resultsreinforce the association between HLA-A3101 and carbamazepine hypersensitivity We didnot identify significant genetic associations with MPE among Han Chinese patients

ConclusionsThe identification of genetic predictors of MPE in CFHR4 and CFH members of thecomplement factor Hndashrelated protein family suggest a new link between regulation of thecomplement system alternative pathway and phenytoin-induced hypersensitivity in European-ancestral patients

RELATED ARTICLE

EditorialGenetic testing to preventadverse reactions toantiepileptic drugsPrimum non nocere

Page 155

These authors contributed equally to this work

Author affiliations are provided at the end of the article

Coinvestigators are listed at httplinkslwwcomWNLA153

Go to NeurologyorgN for full disclosures Funding information and disclosures deemed relevant by the authors if any are provided at the end of the article

The Article Processing Charge was funded by the European Commission OpenAIRE2020 project

This is an open access article distributed under the terms of the Creative Commons Attribution License 40 (CC BY) which permits unrestricted use distribution and reproduction in anymedium provided the original work is properly cited

e332 Copyright copy 2017 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology

Idiosyncratic cutaneous adverse drug reactions (cADRs) canhave a genetic predisposition The HLA-B1502 allele isa predictor of carbamazepine-induced Stevens-Johnson syn-drome and toxic epidermal necrolysis (SJSTEN) in indi-viduals of Han Chinese and Southeast Asian descent whilea recent meta-analysis suggests that the allele is also a signifi-cant risk factor for oxcarbazepine- phenytoin- andlamotrigine-induced SJSTEN12 However the associationwith HLA-B1502 does not extend to the milder but morecommon maculopapular exanthema (MPE) phenotype andthe allele is specific to individuals of Asian descent limitingclinical utility across populations34 HLA-A3101 has beenconfirmed as a transethnic risk factor for carbamazepine-induced cADRs with the allele observed across populations ofEuropean Japanese and Korean descent5ndash8 Recently HLA-A2402 has been shown to associate with SJS in Han Chinesepatients irrespective of causal drug studied9 Genetic variationbeyond the major histocompatibility locus has also been as-sociated with cADRs The CYP2C93 allele correlates withphenytoin hypersensitivity in Han Chinese from Taiwan10

with a similar effect reported in a Thai population11 Howevera genome-wide association study (GWAS) of lamotrigine andphenytoin-induced cADRs in Europeans did not detect sig-nificant predictors12 A summary of the associated genetic riskvariants for cADRs in various populations is provided in tablee-1 (linkslwwcomWNLA56)

The EpiPGX Consortium was established to identify geneticmarkers of epilepsy treatment response The InternationalLeague Against Epilepsy Complex Genetics Consortium(ILAE-CGC) facilitates the discovery of genetic variantsinfluencing epilepsy predisposition13 The EPIGEN consor-tium is a worldwide epilepsy genetics research framework andthe Canadian Pharmacogenomics Network for Drug Safety(CPNDS) is an active surveillance network focused onidentifying genomic markers of severe adverse drug reactions(ADRs) in children and adults14 Collaboration among theseconsortia has provided detailed phenotypes and genotypes forover 15000 epilepsy cases for the investigation of antiepi-leptic drug (AED)ndashinduced MPE

Availing of the joint resources of EpiPGX ILAE-CGCEPIGEN and CPNDS this study aimed to characterizeamong European and Han Chinese populations the geneticpredictors of MPE a cutaneous ADR common to particularAEDs Specifically we set out to test the following hypothe-ses (1) whether population-specific genetic variants predict

MPE (2) whether transethnic genetic variants predict MPE(3) whether population-specific genetic variants predict AED-specific MPE and (4) whether transethnic genetic variantspredict AED-specific MPE

MethodsStandard protocol approvals registrationsand patient consentsAll study participants provided written informed consent forgenetic analysis Local institutional review boards approvedstudy protocols at each contributing site

Study designWe conducted a retrospective case-control study in individualsof European and Han Chinese ethnicity Participants were ex-posed to carbamazepine lamotrigine phenytoin or oxcarba-zepine Our analyses were structured to test genetic variants forassociation with MPE within and across both of the broadancestral groups through logistic regression of genotype dosageand subsequent meta-analysis of regression coefficients Wetested for association with (1) aromatic AED-induced MPE vscontrols tolerant to at least 3 aromatic AEDs (2)carbamazepine-induced MPE vs carbamazepine-tolerant con-trols (3) lamotrigine-induced MPE vs lamotrigine-tolerantcontrols and (4) phenytoin-induced MPE vs phenytoin-tolerant controls Due to small sample size oxcarbazepine-related MPE was not analyzed as an individual case cohort

Cohorts and phenotype definitionEpilepsy cohorts from the ILAE-CGC EpiPGX and EPIGENConsortia were included in the discovery GWAS meta-analysis (table 1) A European-descent replication cohort wasassembled from sites in Brazil Canada (via CPNDS) Liver-pool and other sites across the United Kingdom Cases weredefined as having MPE attributed to carbamazepine lamo-trigine phenytoin or oxcarbazepine as determined by theirclinician occurring within 3months of initiation and resolvingupon dose reduction or AED withdrawal Control patientstrialed carbamazepine lamotrigine phenytoin or oxcarbaze-pine for at least 3months without reporting a cADR Epilepsy-specific patient demographics are presented in table e-2 (linkslwwcomWNLA56)

Genotyping and imputationGenotyping of a subset of EpiPGX samples was performed atdeCODE Genetics (Reykjavik Iceland) using Illumina (San

GlossaryADR = adverse drug reaction AED = antiepileptic drug cADR = cutaneous adverse drug reaction CFH = complement factorHCFHR4 = complement factor Hndashrelated 4 geneCI = confidence intervalCPNDS =Canadian Pharmacogenomics Networkfor Drug SafetyGWAS = genome-wide association studyHLA = human leukocyte antigen ILAE-CGC = International LeagueAgainst Epilepsy Complex Genetics Consortium MPE = maculopapular exanthema OR = odds ratio SJSTEN = Stevens-Johnson syndrome and toxic epidermal necrolysis SNP = single nucleotide polymorphism

NeurologyorgN Neurology | Volume 90 Number 4 | January 23 2018 e333

Diego CA) OmniExpress-12 v11 and OmniExpress-24 v11single nucleotide polymorphism (SNP) arrays The re-mainder of samples were genotyped locally using variousIllumina beadchip SNP arrays details of which are publishedelsewhere13 Genotyping and imputation quality control isdescribed in appendix e-1 (linkslwwcomWNLA57)

Study powerWe estimated that the study had 80 power to detect a ge-netic predictor of relative risk (approximated to odds ratio)ge3 with an allele frequency ge2 and an α level of 125 times 10minus8Power for AED-specific and population-specific analyses aredetailed in appendix e-1 (linkslwwcomWNLA57)

Statistical analysesAssociation analyses were conducted within the Europeanand Asian subgroups using an additive logistic regressionmodel To account for genotype uncertainty SNPTEST wasused to apply a missing data likelihood score model thatincluded sex clinical site and 5 principal components ascovariates to control for bias and population stratification15

Fixed effects meta-analyses were conducted across the Eu-ropean and Asian subgroups using the software packageMETAL applying genomic control correction withincohorts16 The threshold for statistical significance was set at125 times 10minus8 reflecting an empirical Bonferroni correction for4 tests of the standard 5 times 10minus8 genome-wide significancethreshold Conditional association analysis was performedon loci containing significant markers to establish whetherother genetic variants in the region (1 Mb upstream anddownstream) were independently associated with MPE Theconditional threshold for significance was set at 5 times 10minus6based on a genome-wide estimation of 10000 imputed var-iants per 2 MB region13 We applied the Stouffer z trend testto the combined results from the discovery and replicationcohorts

Confirmatory genotypingWhere an association signal satisfied the threshold for sig-nificance additional genotyping and resequencing were per-formed in a subset of patients and results were compared withimputation dosage files The variant rs78239784 was

confirmed by Sanger sequencing in 100 patients from theoriginal discovery cohort For the purpose of replication wegenotyped the rs78239784 variant in an independent cohortof 13 phenytoin-induced MPE cases and 88 phenytoin-tolerant controls

ResultsCohort descriptionIn total 375 MPE cases and 1321 controls satisfied our cri-teria for inclusion in the discovery analyses (see Methods andtable 1) There were 16 patients with cross-reactivity to 2 ormore aromatic AEDs 8 of whom were hypersensitive tocarbamazepine and lamotrigine Genome-wide array data for323 cases and 1321 controls were available for analysis BroadEuropean or Han Chinese ancestry was assigned to eachparticipant according to principal components analysis (figuree-1 linkslwwcomWNLA55)

Genome-wide association analysis of broadaromatic AEDndashinduced MPEAfter quality control (see appendix e-1 figure e-2 linkslwwcomWNLA55 for details) 3693290 variants remained foranalysis in the European dataset and 4402554 variants in theHan Chinese dataset We only considered autosomal SNPs inour analyses To test hypothesis (1) that population-specificgenetic markers predispose to MPE a logistic regressionanalysis of all MPE cases was performed separately in theEuropean and Han Chinese ancestral subgroups We did notobserve any genome-wide significant markers for MPE due toany aromatic AED in either Europeans or Han Chinese Thestudy was powered to detect an effect of relative risk gt35 inthe European cohort and gt5 in the Han Chinese

To test hypothesis (2) that transethnic genetic markerspredispose to MPE a fixed-effects meta-analysis of the asso-ciation results for European and Han Chinese ancestral sub-groups was performed We did not observe any genome-widesignificant markers for MPE shared among European or HanChinese subgroups (figure 1A) This analysis was powered todetect an effect size gt3

Table 1 Breakdown of antiepileptic drug (AED)ndashinduced maculopapular exanthema (MPE) cases and AED-tolerantcontrols in discovery dataset

Ethnicity

All aromatic AEDs CBZ LTG PHT

MPEa Controlb MPE Control MPE Control MPE Control

European 259 979 95 869 118 812 52 472

Han Chinesec 116 342 85 197 16 32 22 58

Subtotal 375 1321 180 1066 134 844 74 530

Abbreviations CBZ = carbamazepine ILAE = International League Against Epilepsy LTG = lamotrigine PHT = phenytoina Individual participant counts only despite 16 patients being cross-reactive to more than 1 AEDb A total of 1321 controls were tolerant to all 3 of CBZ LTG and PHTc Fifty-two carbamazepine-induced MPE cases from Guangzhou were available for analysis of human leukocyte antigen serotype data only

e334 Neurology | Volume 90 Number 4 | January 23 2018 NeurologyorgN

Genome-wide association analysis of specificaromatic AEDndashinduced MPETo test hypothesis (3) that genetic variants for MPE areAED-specific and population-specific logistic regressionanalyses of AED-specific MPE was performed separately inthe European and Han Chinese ancestral groups (figures e-3and e-4 linkslwwcomWNLA55) Within the Europeansubgroup HLA-A3101 was significantly associated withcarbamazepine-induced MPE in Europeans (p = 147 times 10minus10odds ratio [OR] [95 confidence interval (CI)] 55[30ndash10]) Conditioning on HLA-A3101 did not revealadditional variants within the human leukocyte antigen(HLA) region that were independently contributing tocarbamazepine-induced MPE No genome-wide significantsignals for lamotrigine-induced MPE were observed inEuropeans This analysis was powered to detect an effectsize gt6

For phenytoin-induced MPE we identified a significant as-sociation with rs78239784 an intronic variant of the com-plement factor Hndashrelated 4 gene (CFHR4) The risk allele Ghad a minor allele frequency of 12 in our Europeanphenytoin-inducedMPE cases compared to 15 in Europeanphenytoin-tolerant controls (p = 294 times 10minus10 OR [95 CI]88 [40ndash19] figure 2) Conditioning on rs78239784 did notreveal additional variants in this locus that were independentlycontributing to phenytoin-induced MPE Using 1000Genomes Phase III European population data rs78239784was found to be in complete linkage disequilibrium withrs35274867 (r2 = 1 and Drsquo = 1) a missense variant coding foran asparagine to tyrosine substitution at amino acid 1050 ofthe complement factor H (CFH) gene The missense variantwas not present in our association results as it was filteredduring quality control because the imputation score waslt095 The imputation accuracy of the top variant rs78239784was confirmed in our cohort via Sanger sequencing andTaqMan approaches Of 100 samples tested a 100 con-cordance rate was found between imputed and resequencedgenotypes Within the Han Chinese subgroup no significantassociations were found between autosomal SNPs or HLAalleles and AED-specific MPE Summary results for knownrisk loci in our dataset were scrutinized and are presented intable 2 None of these loci was even nominally significant (p gt005) in the Han Chinese subgroup

In order to test hypothesis (4) that transethnic geneticmarkers predispose to AED-specific MPE we meta-analyzedp values from association results for carbamazepine lamo-trigine and phenytoin individually across European and HanChinese ancestral subgroups There were no shared genome-wide significant markers among our meta-analyses of AED-specific MPE (figure 1 BndashD)

Replication of CFHR4 signalTo replicate the association with phenytoin-induced MPE inan independent cohort the variant rs78239784 was geno-typed in self-reported European-descent cases and controls

recruited through centers in Liverpool (United Kingdom)Sao Paolo (Brazil) and across Canada (CPNDS) Two het-erozygous carriers were identified among 13 phenytoin-induced MPE cases while only a single carrier was observedamong 88 phenytoin-tolerant controls yielding a 2-tailedFisher exact p value of 0044 Pooling all cases and controlstogether we report an overall p value of 45 times 10minus11 (witha combined OR [95 CI] 7 [32ndash16]) for the associationbetween rs78239784 and phenytoin-induced MPE inEuropeans (table 3)

DiscussionWe detected a strong association between variants in thecomplement factor H regulatory pathway and phenytoin-induced MPE in a European-descent patient population Thepresence of the associated genotype increases risk for MPE 6-fold Our results indicate that risk variants for MPE tend to bedrug-specific and population-specific

These results point to the regulators of complement activa-tion gene cluster as a genetic locus contributing to the onset ofhypersensitivity to phenytoin The most significant variant inour European subgroup analysis rs78239784 (c59-2448TgtG) tags the missense variant rs35274867 (pN1050Y) in CFH suggesting aberrant complement activationas a potential causal mechanism in a subset of phenytoin-sensitive individuals According to data from the Exome Ag-gregation Consortium CFH N1050Y has an allele frequencyof approximately 2 in Europeans 3 in African sub-populations less than 1 in South Asians and is almost in-variant in East Asians17 Given the absence of this allele in EastAsian populations the lack of an association between theCFHlocus and MPE in our Han Chinese cohort is unsurprisingWe propose that population-specific independent rare var-iants of large effect may explain a proportion of MPE casesa similar paradigm to the rare variant model demonstrated inCrohn disease and ulcerative colitis18 CFH N1050Y haspreviously been associated with type 2 diabetesndashassociatedend-stage kidney disease in an African American cohort19

Defects in CFH-related proteins are also associated withoveractivation of the complement immune system and canlead to atypical hemolytic-uremic syndrome C3 glomerul-opathy basal laminar drusen immunoglobulin A nephropa-thies and systemic lupus erythematosus20 Further geneticvariants in CFHR4 and CFH are associated with risk for age-related macular degeneration21 We did not detect thesesymptoms among N1050Y carriers While it is unclearwhether phenytoin directly interacts with circulating CFH-related proteins it does not specifically increase serum com-plement levels22 Our findings offer an expanded insight intothe role of the complement alternative pathway in hyper-sensitivity to AEDs

Phenytoin is still used as a first-line treatment for epilepsy inmany settings and is listed on the WHO list of essential

NeurologyorgN Neurology | Volume 90 Number 4 | January 23 2018 e335

medicines23 Epidemiologic data on prescriptions of AEDs forepilepsy in the United Kingdom show that in 2008 phenyt-oin accounted for 18 of all treated person-years in epilepsyand was most frequently used in the elderly24 Therefore

a clinically useful prognostic test for phenytoin-induced cu-taneous ADRs in European-ancestral individuals would bewelcome The sensitivity of theCFHR4 variant as a prognosticmarker is 16 and the specificity is 97 which corresponds to

Figure 1 Meta-analysis results across European and Han Chinese cohorts

Manhattan (a) and quantilendashquantile(b) plots for the meta-analyses ofmaculopapular exanthema vs tolerantcontrols for (A) any antiepileptic drug(genomic inflation factor [λ] = 101) (B)carbamazepine (λ = 101) (C) lamo-trigine (λ = 099) and (D) phenytoin (λ =098)

e336 Neurology | Volume 90 Number 4 | January 23 2018 NeurologyorgN

a positive likelihood ratio of 593 (95 CI 28ndash126) anda negative likelihood ratio of 086 (95 CI 08ndash09) As-suming the pretest probability of the ADR is 5 a positivetest for this marker increases the probability of MPE tophenytoin sixfold to 30 while a negative test reduces theprobability marginally to 43 There are an estimated 6million people with epilepsy in Europe which means ap-proximately 90000 people are at-risk carriers of this muta-tion25 We estimate that 208 (95 CI 103ndash431) patients ofEuropean ancestry would need to be screened to prevent onecase based on a previously reported formula26 which corre-sponds to an absolute risk reduction of 0005 (95 CI0002ndash0009) As a comparison it is estimated that 442 Han

Chinese patients would be needed to screened forHLA-B1502 in order to prevent a single carbamazepine-induced SJSTEN case We would suggest that the clinical utility and cost-effectiveness of implementing preemptive screening be eval-uated through a prospective study

We did not replicate the association between CYP2C93(rs1057910) and MPE in our cohort irrespective of ethnicityor AED This is not surprising given that the original associ-ation with phenytoin in Han Chinese was largely driven bySJSTEN cases which were excluded from this analysis Wedid however detect a nonsignificant enrichment ofCYP2C93 (p = 008) among the European phenytoin-

Figure 2 Intronic CFHR4 variants are associated with phenytoin-induced maculopapular exanthema (MPE) in Europeans

(A) Manhattan and (B) quantilendashquantile plot of phenytoin-induced MPE in the European subgroup (λ = 101) The LocusZoom plot (C) highlights the mostsignificant single nucleotide polymorphism (SNP) rs78239784 (purple dot) is an intronic variant in CFHR4

NeurologyorgN Neurology | Volume 90 Number 4 | January 23 2018 e337

induced MPE cases but the effect size we observe (OR 18) issmaller than previously reported for phenytoin-induced MPEin Han Chinese (OR 55)10 Notably 2 CYP2C93 carriersamong the phenytoin-induced MPE cases were also hetero-zygous for the CFHR4 variant while only 3 of 560 controlswere jointly heterozygous The frequency of CYP2C93 dif-fers between controls from European and Han Chinese sub-groups in our study which is in accordance with backgroundpopulation frequency reported in the Exome AggregationConsortium (European 7 East Asian 3) While ourresults do not support a significant effect of CYP2C93 inMPE larger cohorts including severe cADR cases may resolvethe extent of the association across populations

HLA-A3101 was the most strongly associated marker withcarbamazepine-induced MPE in Europeans in this studyForty-three of the 95 cases studied here were also included inthe discovery publication8 but the effect of the allele remainssignificant when we restrict to new cases only (p = 4 times 10minus7)thus providing an additional independent replication of theinitial finding We confirm thatHLA-B1502 is not associatedwith carbamazepine-induced MPE in Han Chinese and nonovel signals emerged for carbamazepine-induced MPE ineither population No significant predictors of lamotrigine-induced MPE were observed in either population testedHLA-A2402 was not significantly associated withlamotrigine-induced MPE in either of the European or HanChinese ancestral subgroups rather this allele was observedto be more frequent among our lamotrigine-tolerant controls

Our meta-analyses did not reveal any significant transethnicgenetic markers for MPE due to any AED There were con-siderably more European-descent patients in this analysis thanany other ethnicity and we recognize this as a limitation of thestudy Analysis of non-European cohorts is warranted Asecond limitation of our study was the low number of HanChinese lamotrigine-related MPE cases relative to European-descent cases Therefore we cannot conclusively rule outgenetic predictors of modest effect size for MPE to lamo-trigine in Han Chinese or other non-European descent pop-ulations We recognize that our replication cohort forphenytoin is small and comprises self-reported ancestralEuropeans Since we did not have full genotype array data forthese individuals we relied on Fisher exact test for calculatingsignificance rather than logistic regression with correction forprincipal components Additional studies of larger sample sizeare required to further characterize the association improvethe estimation of the risk effect size and determine theprognostic ability and economics of screening for this markerFinally as this study was not powered to investigate MPEattributed to oxcarbazepine due to low sample size furtherinvestigation is warranted

We have identified a genetic predictor for a common adversereaction to phenytoin in European-descent patients addinga new pharmacogenetic marker for potential use in thetreatment of epilepsy This finding adds to the list of geneticTa

ble

2Associationtest

resu

ltsforrisk

allelesformac

ulopap

ularex

anthem

a(M

PE)

across

ethnicities

Dru

gMark

er

Euro

pean

HanChinese

MPE

Control

OR(95

CI)

pValue

MPE

Control

OR(95

CI)

pValue

Homozy

gosity

hete

rozy

gosity

refere

nce

(MAF)

Homozy

gosity

hete

rozy

gosity

refere

nce

(MAF)

Homozy

gosity

hete

rozy

gosity

refere

nce

(MAF)

Homozy

gosity

hete

rozy

gosity

refere

nce

(MAF)

CBZ

HLA

-B1

502

0095

(0)

0086

9(0)

mdashmdash

0221

(004)

322

162

(007)

06(01ndash28)

053

CBZ

HLA

-A3

101

016

79(008)

027

841

(002)

55(30ndash10

)147

times10

minus10

0122

(002)

010

156

(003)

08(01ndash68)

081

LTG

HLA

-A2

402

214

102

(008)

812

268

1(009)

09(05ndash15)

071

0412

(012)

1720

(016)

07(01ndash35)

062

PHT

CYP

2C9

3010

42(01)

051

421

(005)

18(09ndash38)

008

0121

(002)

0256

(002)

09(01ndash12

)092

PHT

rs78

2397

842842

(012)

014

458

(002)

88(40ndash19

)294

times10

minus10

0022

(0)

0058

(0)

mdashmdash

Abbreviations

CBZ=ca

rbam

azep

ine

CI=

confiden

ceintervalL

TG=lamotrigine

MAF=mea

nallele

freq

uen

cyO

R=oddsratioP

HT=phen

ytoin

Rep

rese

ntative

risk

allelesfrom

previousge

nome-wideas

sociationstudyfindings

andthisreportforan

tiep

ilepticdru

gndashinduce

dskin

rash

studies(CBZ

LTGandPHT)

wereas

sessed

foras

sociationin

ourstudypValues

from

logisticregres

sionmodel

withse

xan

d5principal

componen

tsas

cova

riates

e338 Neurology | Volume 90 Number 4 | January 23 2018 NeurologyorgN

predictors of hypersensitivity to anticonvulsant therapy andopens up a new avenue for understanding the biology un-derlying cutaneous adverse reactions This finding can ad-vance genetic testing in the clinic as it expands the array ofgenetic tests available to aid clinicians in reducing overall ratesof discontinuation due to adverse events and improving pa-tient safety

AffiliationsFrom Molecular and Cellular Therapeutics (MM NDGLC) and FutureNeuro Research Centre (GLC) RoyalCollege of Surgeons in Ireland Dublin Centre for GenomicSciences (HG SC LB) Li Ka Shing Faculty of MedicineUniversity of Hong Kong deCODE GeneticsAmgen Inc(AI LJG KS) Reykjavik Iceland UCL Genetics Institute(DS) University College London UK Centre forMolecularMedicine and Therapeutics BC Childrenrsquos Hospital ResearchInstitute Faculty of Medicine (GEBW) Departments ofPediatrics and Medical Genetics (CJDR) and Division ofTranslational Therapeutics Department of Pediatrics (BCC) University of British Columbia Vancouver CanadaDepartment of Molecular and Clinical Pharmacology (JEZGJS AGM PA MP) Institute of Translational Medi-cine University of Liverpool UK Department of MedicalGenetics (RS IL-C) School of Medical Sciences Univer-sity of Campinas and the Brazilian Institute of Neuroscienceand Neurotechnology (BRAINN) Department of Neurology(CY FC) School of Medical Sciences University ofCampinas and the Brazilian Institute of Neuroscience andNeurotechnology (BRAINN) Campinas SP Brazil HertieInstitute for Clinical Brain Research (MK HL) Universityof Tubingen Germany Department of Clinical and Experi-mental Epilepsy (SW FB SR AA KH CL SMS) andNIHR University College London Hospitals Biomedical Re-search Centre (JWS) UCL Institute of Neurology LondonUK Laboratory of Experimental Neurology (CD) HopitalErasme Universite Libre de Bruxelles Brussels Belgium TheWalton Centre NHS Foundation Trust (AGM PA) Liv-erpool Epilepsy Unit (MJB) West Glasgow ACH-YorkhillDepartment of Biostatistics (BF) Institute of TranslationalMedicine University of Liverpool Division of Brain Sciences(MRJ) Imperial College Faculty of Medicine London UK

Division of Neurosciences (BPCK) and Center for Mo-lecular Medicine (MM) Universitair Medisch CentrumUtrecht Netherlands Pediatric Neurology and MuscularDiseases Unit Department of Neurosciences RehabilitationOphthalmology Genetics Maternal and Child Health Uni-versity of Genoa (PS) and Laboratory of Neurogenetics andNeuroscience (FZ) Institute G Gaslini Department ofNeuroscience (AC) Reproductive and Odontostomato-logical Sciences Federico II University Naples Departmentof Epileptology (WSK) University of Bonn GermanyStichting Epilepsie Instellingen Nederland (SEIN) (JWS)Heemstede Netherlands The Chalfont Centre for Epilepsy(JWS SMS) Epilepsy Society Chalfont St PetersBuckinghamshire UK Epilepsy Center Frankfurt Rhine-Main (KMK) Department of Neurology Center ofNeurology and Neurosurgery University Hospital Goethe-University Frankfurt Epilepsy Center Hessen (KMK)Department of Neurology University Hospitals Giessen ampMarburg and Philipps-University Marburg Germany Neu-rogenetics Group (SW) VIB-UAntwerp Center for Mo-lecular Neurology Neurology Department (SW) UniversityHospital Antwerp BelgiumDepartment of Neurology (MK)Medical University of Vienna Austria Faculty of Medicine(KS) University of Iceland Reykjavik Luxembourg Centrefor Systems Biomedicine (RK) University of LuxembourgUniversity of Toronto (NS) Institute of Neuroscience andThe Second Affiliated Hospital of Guangzhou Medical Uni-versity (W-pL) Key Laboratory of Neurogenetics andChannelopathies of Guangdong Province and the Ministry ofEducation of China China and Departments of Medicine andNeurology (TJO PK) University of Melbourne RoyalMelbourne Hospital Australia

Author contributionsMMC HG LB SC PK and GLC contributed to theconception and design of the study and the acquisitionanalysis and interpretation of data MMC and HG per-formed all statistical analysis AI SMS RS SW FB SRKH CL AGM PA MJB BF MRJ NS GEBWCJD BCC DS JEZ MK MP AA CD GJS BPCK PS FZ AC WSK JWS HL KMK SW MK

Table 3 rs78239784 associates with phenytoin-induced maculopapular exanthema in Europeans

rs78239784 Discovery GWAS Replication Combined Control

N 52 13 63 560

Frequency (G) 012 008 011 001

p Value 29 times 10minus10a 0044b 45 times 10minus11 mdash

OR (95 CI) 88 (40ndash19) 15 (13ndash17) 70 (32ndash16) mdash

Abbreviations CI = confidence interval GWAS = genome-wide association study OR = odds ratioDiscovery replication and combined association test results of rs78239784 with phenytoin-induced cutaneous adverse drug reactions in Europeansa p Value from logistic regression model with sex and 5 principal components as covariatesb p Value from Fisher exact test

NeurologyorgN Neurology | Volume 90 Number 4 | January 23 2018 e339

LJG RK ND CY FC IL-C TJO W-pL LJGand KS contributed to the acquisition analysis and in-terpretation of data All authors contributed to the criticalrevision of the final version of the manuscript for importantintellectual content

AcknowledgmentThe authors thank the following coinvestigators the membersof the EpiPGX Consortium the members of the CanadianPharmacogenomics Network for Drug Safety (CPNDS)Consortium the members of the International League AgainstEpilepsy Consortium on Complex Epilepsies (ILAE-CGC)and the members of the EPIGEN Consortium (a full list ofcoinvestigators is found in the e-Appendix) Ana Fulgenico-Maisch Simona Donatello Clare OrsquoKennedy Lisa SlatteryPaula Corr and Joanna Fay for clinical assessment of patientstechnical assistance with DNA extraction and sample manage-ment and the patients for their participation in the study

Study fundingThis study was not industry-sponsored The work was sup-ported by a grant from the European Commission (7thFramework Programme Grant 279062 EpiPGX) MMCand GLC are supported by Science Foundation Irelandgrant 13CDA2223 and an RCSI seed funding grant GA 14-1899 This project was supported by the General ResearchFunds (HKU762308M and HKU774707M to SCCUHK446606M to PK) and Health and Medical ResearchFund (HMRF 01120086 to PK) from Hong Kong Someresults presented in this article were prepared using the HPCfacilities of the University of Luxembourg This work waspartly undertaken at UCLHUCL which received a pro-portion of funding from the Department of Healthrsquos NIHRBiomedical Research Centres funding scheme (JWS SMS)The work was also supported by the Epilepsy Society UK(JWS SMS) by the foundation ldquono epileprdquo the GermanChapter of the ILAE (DGfE) (both to HL) FC and IL-Care supported by Fundaccedilatildeo de Amparo a Pesquisa do Estadode Satildeo Paulo Brazil through grant 201307559-3 JEZ andMP thank the NHS Chair of Pharmacogenetics programmeand MRC Centre for Drug Safety Science for support inLiverpool BCC and CJDR are supported by the CanadianInstitutes of Health Research (CIHR) Drug Safety and Ef-fectiveness Network (FRN-117588) the Canada Foundationfor Innovation and the Canadian Dermatology FoundationGEBW is supported by a CIHR Fellowship The funders ofthe study had no role in the study design data collection dataanalysis data interpretation or writing of the report MMCHG and GLC had full access to all the data in the study andthe corresponding authors had final responsibility for thedecision to submit for publication

DisclosureM McCormack and H Gui report no disclosures relevant tothe manuscript A Ingason is an employee of deCODEGeneticsAmgen D Speed G Wright E Zhang R SecolinC Yasuda M Kwok S Wolking F Becker and S Rau report

no disclosures relevant to the manuscript A Avbersek isemployed by UCB Pharma SPRL Belgium as Associate Di-rector K Heggeli C Leu C Depondt and G Sills report nodisclosures relevant to the manuscript A Marson was awar-ded grants from GSK Eisai and UCB Pharma which fundedthe National Audit of Seizure Management in HospitalsP Auce M Brodie B Francis M Johnson B KoelemanP Striano A Coppola F Zara and W Kunz report no dis-closures relevant to the manuscript J Sander has served onscientific advisory boards for UCB Pharma and Eisai hasserved on speakerrsquos bureaus for UCB Pharma Eisai Teva andLundbeck has received research support from UCB PharmaGSK Eisai The Marvin Weil Epilepsy Research Fund andNL Nationaal Epilepsie Fonds and his current position isendowed by the UK Epilepsy Society H Lerche has receivedspeaker or consultancy fees or travel support from BialDesitin Eisai GlaxoSmithKline Pfizer UCB Pharma orValeant K Klein reports personal fees from UCB PharmaNovartis Pharma AG and Eisai outside of the submittedwork AC was awarded a grant from EISAI and personal feesfor speaking from Eisai outside of the submitted workS Weckhuysen and M Krenn report no disclosures relevantto the manuscript L Gudmundsson is an employee of de-CODE GeneticsAmgen K Stefansson is an employee ofdeCODE GeneticsAmgen R Krause N Shear C Ross andN Delanty report no disclosures relevant to the manuscriptM Pirmohamed reports grants fromMRC and grants from UKDepartment of Health during the conduct of the studyB Carleton through the Pharmaceutical Outcomes Programme(POPi) has received financial support for its pharmacoge-nomics research from the Canada Foundation for Innovation(CFI) Canadian Institutes of Health Research Genome Can-ada Genome British Columbia and the Provincial HealthServices Authority Prof Carelton has also received support bythe University of British Columbia Child amp Family ResearchInstitute Vancouver and Pfizer Prof Carleton has a patent andapplications pending for biomarkers of anthracycline-inducedcardiotoxicity and cisplatin-induced ototoxicity but no relevantfinancial disclosures relevant to this study F Cendes reportsspeaking fees from UCB Pharma outside of the current workI Cendes-Lopes and W Liao report no disclosures relevant tothe manuscript T OrsquoBrien reports grants from The RoyalMelbourne Foundation during the conduct of the studyS Sisodiya and S Cherny report no disclosures relevant to themanuscript P Kwan has received speaker or consultancy feesandor research grants from Eisai GlaxoSmithKline JohnsonampJohnson Pfizer and UCB Pharma L Baum and G Cavallerireport no disclosures relevant to the manuscript Go toNeurologyorgN for full disclosures

Received February 27 2017 Accepted in final form October 2 2017

References1 Li X Yu K Mei S et al HLA-B1502 increases the risk of phenytoin or lamotrigine

induced Stevens-Johnson syndrometoxic epidermal necrolysis evidence froma meta-analysis of nine case-control studies Drug Res 201565107ndash111

2 ChungWH Hung SI Hong HS et al Medical genetics a marker for Stevens-Johnsonsyndrome Nature 2004428486

3 ManCBKwan P BaumL et al Association betweenHLA-B1502 allele and antiepilepticdrug-induced cutaneous reactions in Han Chinese Epilepsia 2007481015ndash1018

e340 Neurology | Volume 90 Number 4 | January 23 2018 NeurologyorgN

4 Lonjou C Thomas L Borot N et al A marker for Stevens-Johnson syndromeethnicity matters Pharmacogenomics J 20066265ndash268

5 Amstutz U Ross CJ Castro-Pastrana LI et al HLA-A 3101 and HLA-B 1502 asgenetic markers for carbamazepine hypersensitivity in children Clin Pharmacol Ther201394142ndash149

6 Kim SH Lee KW Song WJ et al Carbamazepine-induced severe cutaneous adversereactions and HLA genotypes in Koreans Epilepsy Res 201197190ndash197

7 Ozeki T Mushiroda T Yowang A et al Genome-wide association study identifiesHLA-A3101 allele as a genetic risk factor for carbamazepine-induced cutaneousadverse drug reactions in Japanese population Hum Mol Genet 2011201034ndash1041

8 McCormack M Alfirevic A Bourgeois S et al HLA-A3101 and carbamazepine-induced hypersensitivity reactions in Europeans N Engl J Med 20113641134ndash1143

9 Shi YWMin FL Zhou D et al HLA-A2402 as a common risk factor for antiepilepticdrug-induced cutaneous adverse reactions Neurology 2017882183ndash2191

10 Chung WH Chang WC Lee YS et al Genetic variants associated with phenytoin-related severe cutaneous adverse reactions JAMA 2014312525ndash534

11 Tassaneeyakul W Prabmeechai N Sukasem C et al Associations between HLA classI and cytochrome P450 2C9 genetic polymorphisms and phenytoin-related severecutaneous adverse reactions in a Thai population Pharmacogenet Genomics 201626225ndash234

12 McCormack M Urban TJ Shianna KV et al Genome-wide mapping for clinicallyrelevant predictors of lamotrigine- and phenytoin-induced hypersensitivity reactionsPharmacogenomics 201213399ndash405

13 International League Against Epilepsy Consortium on Complex Epilepsies Geneticdeterminants of common epilepsies a meta-analysis of genome-wide associationstudies Lancet Neurol 201413893ndash903

14 Carleton B Poole R Smith M et al Adverse drug reaction active surveillance de-veloping a national network in Canadarsquos childrenrsquos hospitals PharmacoepidemiolDrug Saf 200918713ndash721

15 Marchini J Howie B Myers S McVean G Donnelly P A new multipoint method forgenome-wide association studies by imputation of genotypes Nat Genet 200739906ndash913

16 Willer CJ Li Y Abecasis GR METAL fast and efficient meta-analysis of genomewideassociation scans Bioinformatics 2010262190ndash2191

17 Lek M Karczewski KJ Minikel EV et al Analysis of protein-coding genetic variationin 60706 humans Nature 2016536285ndash291

18 Beaudoin M Goyette P Boucher G et al Deep resequencing of GWAS loci identifiesrare variants in CARD9 IL23R and RNF186 that are associated with ulcerative colitisPLoS Genet 20139e1003723

19 Bonomo JA Palmer ND Hicks PJ et al Complement factor H gene associations withend-stage kidney disease in African Americans Nephrol Dial Transpl 2014291409ndash1414

20 Skerka C Chen Q Fremeaux-Bacchi V Roumenina LT Complement factor H re-lated proteins (CFHRs) Mol Immunol 201356170ndash180

21 Fritsche LG Igl W Bailey JN et al A large genome-wide association study of age-related macular degeneration highlights contributions of rare and common variantsNat Genet 201648134ndash143

22 Basaran N Kansu E Hincal F Serum immunoglobulins complement levelsand lymphocyte subpopulations in phenytoin-treated epileptic patients Immuno-pharmacol Immunotoxicol 198911335ndash346

23 The selection and use of essential medicines World Health Organ Tech Rep Ser2015 viindashxv1ndash546

24 Nicholas JM Ridsdale L Richardson MP Ashworth M Gulliford MC Trends inantiepileptic drug utilisation in UK primary care 1993-2008 cohort study using theGeneral Practice Research Database Seizure 201221466ndash470

25 BaulacM de Boer H Elger C et al Epilepsy priorities in Europe a report of the ILAE-IBE Epilepsy Advocacy Europe Task Force Epilepsia 2015561687ndash1695

26 Chen Z Liew D Kwan P Real-world efficiency of pharmacogenetic screening forcarbamazepine-induced severe cutaneous adverse reactions PLoS One 20149e96990

NeurologyorgN Neurology | Volume 90 Number 4 | January 23 2018 e341

SOURCE ARTICLE NPubcomi280xn

Genetic variation in CFH predictsphenytoin-induced maculopapular exanthemain European-descent patientsMark McCormack PhD Honsheng Gui PhD Andres Ingason PhD Doug Speed PhD Galen EB Wright PhD J Eunice Zhang PhD

Rodrigo Secolin PhD Clarissa Yasuda MD PhD Maxwell Kwok MPhil StefanWolking MD Felicitas Becker Sarah Rau MPH

Andreja Avbersek MD Kristin Heggeli BSc Costin Leu PhD Chantal Depondt PhD Graeme J Sills PhD Anthony G Marson MD FRCP

Pauls Auce MD Martin J Brodie MD FRCP Ben Francis PhD Michael R Johnson DPhil FRCP Bobby PC Koeleman PhD

PasqualeStrianoMDPhDAntoniettaCoppolaMD FedericoZara PhDWolframSKunz PhD JosemirWSander FRCPHolger LercheMD

Karl Martin Klein MD PhD SarahWeckhuysen MD PhD Martin Krenn MD Larus J Gudmundsson PhD Kari Stefansson PhD

Roland Krause PhD Neil Shear MD Colin JD Ross PhD Norman Delanty FRCP for the EPIGEN Consortium Munir Pirmohamed FRCP

Bruce C Carleton PharmD for the Canadian Pharmacogenomics Network for Drug Safety Fernando Cendes MD PhD

Iscia Lopes-Cendes MD PhD Wei-ping Liao MD PhD Terence J OrsquoBrien PhD Sanjay M Sisodiya FRCP for the EpiPGX Consortium

Stacey Cherny PhD Patrick Kwan PhD Larry Baum PhD for the International League Against Epilepsy Consortium on Complex Epilepsies

and Gianpiero L Cavalleri PhD

Cite as Neurologyreg 201890e332-e341 doi101212WNL0000000000004853

Correspondence

Prof Cavalleri

gcavallerircsiie

Study questionAre there genetic predictors of maculopapular exanthema (MPE)a cutaneous adverse drug reaction common to antiepileptic drugs(AEDs) in European and Han Chinese populations

Summary answerAn association was found between a rare variant in the CFHR4 gene andphenytoin-induced MPE in cohorts of European descent

What is known and what this article addsSeveral genetic predictors of hypersensitivity to anticonvulsanttherapy exist The study identified a new pharmacogenetics markerfor potential use in the treatment of epilepsy

Participants and settingThe study was a case-control genome-wide association study (GWAS)of autosomal genotypes including Class I and II human leukocyte an-tigen (HLA) alleles in 323 cases and 1321 drug-tolerant controls fromepilepsy cohorts of northern European and Han Chinese descent Theanalyses tested genetic variants for association with MPE within andacross both of the broad ancestral groups Associations were testedas follows (1) aromatic AED-induced MPE vs controls tolerant toat least 3 aromatic AEDs (2) carbamazepine-induced MPEvs carbamazepine-tolerant controls (3) lamotrigine-induced MPEvs lamotrigine-tolerant controls and (4) phenytoin-inducedMPE vsphenytoin-tolerant controls

Design size and durationThe study is a retrospective casendashcontrol study in individuals of Euro-pean and Han Chinese ethnicity in which participants were exposed tocarbamazepine lamotrigine phenytoin or oxcarbazepine

Primary outcomesThere is a strong association between variants in the CFHR4 and CFHgenes and phenytoin-induced MPE in a European-descent patientpopulation

Main results and the role of chanceAn association was noted between a rare variant in theCFHR4 gene andphenytoin-inducedMPE inEuropeans (p= 45 times 10minus11 odds ratio [95confidence interval] = 7 [32ndash16]) This variant is in complete linkage

disequilibriumwith amissense variant (N1050Y) in theCFH gene Alsoan association between HLA-A3101 and carbamazepine hypersensi-tivity was reinforced

Bias confounding and other reasons for cautionThe study analyzed considerably more European-descent patientsand a relatively low number of Han Chinese lamotrigine-relatedMPE cases Overall the small sample sizes limit conclusions innonndashEuropean descent populations

Generalizability to other populationsThe small size and predominantly European origin of the studysample and low numbers of Han Chinese participants may limitgeneralizability to other ethnic groups

Study fundingpotential competing interestsThe study was funded by a group of foundation government anduniversity grants Go to NeurologyorgN for full disclosures

A draft of the short-form article was written by E Feric a writer with Editage a division of Cactus Communications The authors of thefull-length article and the journal editors edited and approved the final version

Copyright copy 2018 American Academy of Neurology 165

SHORT-FORM ARTICLE

DOI 101212WNL0000000000004853201890e332-e341 Published Online before print December 29 2017Neurology Mark McCormack Hongsheng Gui Andreacutes Ingason et al

European-descent patients predicts phenytoin-induced maculopapular exanthema inCFHGenetic variation in

This information is current as of December 29 2017

ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2017 The Author(s) Published by

reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology

ServicesUpdated Information amp

httpnneurologyorgcontent904e332fullincluding high resolution figures can be found at

References httpnneurologyorgcontent904e332fullref-list-1

This article cites 25 articles 1 of which you can access for free at

Citations httpnneurologyorgcontent904e332fullotherarticles

This article has been cited by 1 HighWire-hosted articles

Subspecialty Collections

httpnneurologyorgcgicollectionpublic_healthPublic health

httpnneurologyorgcgicollectioncase_control_studiesCase control studies

httpnneurologyorgcgicollectionassociation_studies_in_geneticsAssociation studies in genetics

httpnneurologyorgcgicollectionantiepileptic_drugsAntiepileptic drugs

httpnneurologyorgcgicollectionall_global_neurologyAll global neurology

httpnneurologyorgcgicollectionall_epilepsy_seizuresAll EpilepsySeizuresfollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online

ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2017 The Author(s) Published by

reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology

ServicesUpdated Information amp

httpnneurologyorgcontent904e332fullincluding high resolution figures can be found at

References httpnneurologyorgcontent904e332fullref-list-1

This article cites 25 articles 1 of which you can access for free at

Citations httpnneurologyorgcontent904e332fullotherarticles

This article has been cited by 1 HighWire-hosted articles

Subspecialty Collections

httpnneurologyorgcgicollectionpublic_healthPublic health

httpnneurologyorgcgicollectioncase_control_studiesCase control studies

httpnneurologyorgcgicollectionassociation_studies_in_geneticsAssociation studies in genetics

httpnneurologyorgcgicollectionantiepileptic_drugsAntiepileptic drugs

httpnneurologyorgcgicollectionall_global_neurologyAll global neurology

httpnneurologyorgcgicollectionall_epilepsy_seizuresAll EpilepsySeizuresfollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpwwwneurologyorgaboutabout_the_journalpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnneurologyorgsubscribersadvertiseInformation about ordering reprints can be found online

ISSN 0028-3878 Online ISSN 1526-632XWolters Kluwer Health Inc on behalf of the American Academy of Neurology All rights reserved Print1951 it is now a weekly with 48 issues per year Copyright Copyright copy 2017 The Author(s) Published by

reg is the official journal of the American Academy of Neurology Published continuously sinceNeurology