HPE Nongenetic Risk Factors

Embed Size (px)

Citation preview

  • 7/29/2019 HPE Nongenetic Risk Factors

    1/13

    American Journal of Medical Genetics Part C (Seminars in Medical Genetics) 154C:7385 (2010)

    A R T I C L E

    Non-Genetic Risk Factors for HoloprosencephalyCANDICE Y. JOHNSON AND SONJA A. RASMUSSEN*

    Holoprosencephaly (HPE) is a congenital defect of the brain characterized by incomplete cleavage of theembryonic forebrain into left and right hemispheres. Although a substantial proportion of cases of HPE can beattributed to genetic abnormalities, the etiology in many cases remains unknown, with non-genetic risk factorsbelieved to be important contributors. Due to the low birth prevalence of this defect, it has proven difficultto conduct studies of sufficient size to identify risk factors with certainty. This article provides a summary of non-genetic risk factors for HPE that have been investigated in case reports and case series, animal studies, andepidemiologic studies, including maternal illnesses, therapeutic and non-therapeutic exposures, nutritionalfactors, and sociodemographic factors. The article also highlights challenges in study design andfurther areas forresearch to better understand the etiology of HPE. Published 2010 Wiley-Liss, Inc.{

    KEY WORDS: holoprosencephaly; risk factors; teratogens; diabetes; salicylates; cholesterol

    How to cite this article: Johnson CY, Rasmussen SA. 2010. Non-genetic risk factors forholoprosencephaly. Am J Med Genet Part C Semin Med Genet 154C:7385.

    INTRODUCTION

    Holoprosencephaly (HPE) is a congen-

    ital defect of the brain characterized by

    incomplete cleavage of the embryonic

    forebraininto left andr ight hemispheres.

    Affected persons often have accompany-

    ing facial dysmorphology, ranging

    from severe (e.g., cyclopia) to more mild

    (e.g., mild hypotelorism or a single

    central maxillary incisor) abnormalities,

    although 10 20% of affected persons

    have no facial abnormalities [Cohen,

    1989a]. The prevalence at birth of

    this condition is about 1 per 10,000

    live births [Croen et al., 1996;

    The prevalence at birth of this

    condition is about 1 per

    10,000 live births.

    Rasmussen et al., 1996]. However, the

    defect appears to occur much more

    frequently earlier in pregnancy, with an

    estimated rate of 4 per 1,000 conceptuses

    [Matsunaga and Shiota, 1977], whichsuggests a high rate of fetal loss.

    However, the defect

    appears to occur much

    more frequently earlier in

    pregnancy, with an

    estimated rate of 4 per

    1,000 conceptuses, which

    suggests a high rate of

    fetal loss.

    In most cases, the causes of HPE

    are unknown. Genetic causes include

    chromosome abnormalities and single

    gene disorders [Munke et al., 1988;Cohen, 2006; Roessler and Muenke,

    2010]. Among persons with HPE, about

    2550% have chromosome abnormal-

    ities and about 1825% have a mutation

    in a single gene that causes syndromic

    HPE [Muenke and Gropman, 2008].

    The most commonly mutated gene in

    persons with HPE is the sonic hedgehog

    (SHH) gene, which codes for the sonic

    hedgehog protein, a secreted protein

    that plays a critical role in patterning of

    the ventral neural tube, the anterior

    posterior limb axis, and ventral somites

    [Muenke and Gropman, 2008]. In

    recent years, much progress has been

    made in understanding genetic factors

    involved in the etiology of HPE

    [Bendavid et al., 2009; Bendavid et al.,

    2010]. However, even among families in

    whom a single gene has been shown

    to increase the risk for HPE, the

    phenotype varies significantly among

    family members who carry the gene,

    The findings and conclusions in this reportare those of the authors and do notnecessarily represent the official position of

    the Centers for Disease Control and Preven-tion.

    Candice Y. Johnson, M.Sc., is a doctoralstudent in the Department of Epidemiologyat Emory University and Guest Researcher atthe National Center on Birth Defects andDevelopmental Disabilities at the Centers forDisease Control and Prevention in Atlanta.She is interested in the epidemiology of birthdefects and the contributions of chronicillness during pregnancy to birth defectetiology.

    Sonja A. Rasmussen, M.D., M.S., is apediatrician and clinical geneticist and cur-rently serves as Senior Scientist at theNational Center on Birth Defects andDevelopmental Disabilities at the Centersfor Disease Control and Prevention inAtlanta. Her research interests include theidentification of risk factors for birth defects,morbidity and mortality associated with birthdefects and genetic conditions, and theimpact of infections on the pregnant womanand her embryo or fetus.

    *Correspondence to: Sonja A. Rasmussen,M.D., M.S., 1600 Clifton Road, MS E-86,CDC, Atlanta, GA 30333.E-mail: [email protected]

    DOI 10.1002/ajmg.c.30242Publishedonline 26 January 2010 in Wiley

    InterScience (www.interscience.wiley.com)

    Published 2010 Wiley-Liss, Inc.{This article is a US Government work and, as such, is inthe public domain in the United States of America.

  • 7/29/2019 HPE Nongenetic Risk Factors

    2/13

    including some family members with a

    normal brain [Roessler et al., 1996;

    Solomon et al., 2010], a finding that

    supports a role for other genetic and

    non-genetic factors [Ming and Muenke,

    2002].

    Evidence that supports the involve-

    ment of certain non-genetic risk factors

    is available from several sources, includ-

    ing case reports/series, experimental

    studies using animal models, and epi-

    demiologic studies. Each of these has an

    important role to play in better under-

    standing non-genetic risk factors for

    HPE, but each has its own strengths

    and limitations [Rasmussen et al., 2007].

    Case reports can alert investigators to

    potential adverse effects of a particular

    exposure so that the exposure can be

    further investigated in an epidemiologicstudy or experimental animal model.

    However, in an individual case report or

    case series, it is difficult to know if the

    adverse effects are caused by the partic-

    ular exposure or if the observed relation

    is coincidental. Studies using an animal

    model allow experimental manipulation

    of the exposure while other factors (such

    as genetic background) are held con-

    stant; however, animal studies may not

    always predict results in humans [Scialli

    et al., 2004]. Epidemiologic studies of

    HPE are challenging, given its low birth

    prevalence, and only a few studies have

    focused on identifying risk factors for

    this defect. These studies have been

    retrospective and have utilized a case

    control study design. The studies are

    often dependent on maternal recall of

    exposures, which can be incomplete and

    possibly biased because of knowledge of

    the outcome (i.e., recall bias) [Kallen,

    2005]. In addition, it may sometimes be

    challenging to determine if a particular

    exposure is the risk factor or whetheranother (confounding) factor may be

    responsible for the observed increased

    risk. In this manuscript, we will review

    theevidence forcertain non-genetic risk

    factors, includingdata from case reports/

    series, animal models, and epidemio-

    logic studies. In addition, we will discuss

    key issues for consideration in future

    studies in order to improve available

    information on non-genetic risk factors

    for HPE.

    EPIDEMIOLOGIC STUDIESOF HOLOPROSENCEPHALY

    We are aware of four casecontrol

    studies that have focused on a spectrum

    of non-genetic risk factors for HPE.

    Because these studies have been used to

    address a large number of risk factors,and thus will be referred to repeatedly

    below, information about them is briefly

    summarized here. In addition to these

    studies, a few other epidemiologic

    studies of specific risk factors (e.g.,

    diabetes and obesity) have also some-

    times included HPE as one of many

    defects studied.

    Human Embryo Center for

    Teratological Studies

    (Kyoto Embryo Study)

    Between 1962 and 1974, a total of

    36,380 conceptuses (resulting in 3,411

    undamaged embryos) were collected by

    Kyoto University in Japan from physi-

    cians performing pregnancy termina-

    tions [Matsunaga and Shiota, 1977].

    Among the 3,411 embryos, 150 with

    HPE were identified. A matched case

    control study was conducted where

    two non-cases were selected from the

    undamaged embryos as controls and

    matched to each HPE case on parityand month of last menstrual period.

    Karyotyping wasnot performed, and the

    proportion of cases with chromosomal

    abnormalities is unknown. Physicians

    performing pregnancy terminations

    were asked to collect information on

    the reproductive and exposure histories

    of the mothers.

    California Birth Defects

    Monitoring Program (CBDMP)Forty-eight cases of non-syndromic

    HPE (live births, fetal deaths

    20 weeks gestation or pregnancy

    terminations) were identified between

    1993 and 1996 using data from the

    CBDMP, a birth defects surveillance

    system that uses active case ascertain-

    ment methods that include review of

    medical records at hospitals and genetics

    clinics in 11 California counties [Croen

    et al., 2000]. Infants with a chromosome

    abnormality were excluded. Controls

    were liveborn infants without birth

    defects randomly selected from births

    in the hospitals in the area. Motherswere

    interviewed by telephone. Another epi-

    demiologic study of HPE using data

    from the CBDMP focused on cases

    identified between 1983 and 1988 and

    included 53 non-syndromic cases

    [Croen et al., 1996].

    Latin American Study of

    Congenital

    Malformations (ECLAMC)

    Between 1967 and2000, 281livebornor

    stillborn (birthweight 500 g) infants

    with non-syndromic HPE were identi-

    fied through the ECLAMC birth defects

    surveillance system, a hospital-basedsystem that includes births in all South

    American countries (except Guyanas)

    [Orioli and Castilla, 2007]. Cases with

    chromosome abnormalities or other

    syndromes were excluded. Controls

    were selected as the next non-mal-

    formed child born in the same hospital

    as the case, matched on sex. Mothers

    were interviewed by the pediatrician

    shortly after the birth.

    National Birth Defects PreventionStudy (NBDPS)

    The NBDPS is an ongoing, population-

    based casecontrol study that ascertains

    liveborn infants, fetal deaths (20

    weeks gestation), and pregnancy termi-

    nations with major birth defects through

    birth defect surveillance systems in

    10 regions of the United States [Yoon

    et al., 2001; Miller et al., 2010]. Cases

    with recognized single gene disorders

    and chromosome abnormalities are

    excluded [Rasmussen et al., 2003].Controls are liveborn infants without

    birth defects selected from study catch-

    ment areas. Mothers of case and control

    infants are interviewed up to 2 years

    after delivery. From 1997 through

    2004, 74 cases of non-syndromic HPE

    were identified [Miller et al., 2010].

    Mothers with pre-existing diabetes were

    excluded from the analysis of HPE

    risk factors. Analyses were adjusted

    for potential confounders including

    74 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) ARTICLE

  • 7/29/2019 HPE Nongenetic Risk Factors

    3/13

    maternal age, education, race/ethnicity,

    number of previous pregnancies, and

    periconceptional use of aspirin and folic

    acid.

    MATERNAL ILLNESSES

    Diabetes

    Maternal diabetes is a well-recognized

    risk factor for birth defects [Becerra

    et al., 1990; Correa et al., 2008] and is

    one of the most extensively studied

    risk factors for HPE. HPE is estimated

    to occur in 1% of births to

    diabetic women [Barr et al., 1983].

    Maternal diabetes is awell-recognized risk factor for

    birth defects and is one of the

    most extensively studied risk

    factors for HPE. HPE is

    estimated to occur in 1% of

    births to diabetic women.

    Surveillance data from the New York

    State Congenital Malformations Regis-

    try and data from a case series collected

    by the Carter Centers for Brain

    Research have noted the prevalence of

    pre-existing diabetes among mothers of

    children with HPE to be 6% and 9%,

    respectively [Olsen et al., 1997; Sta-

    shinko et al., 2004].

    Results from a mouse model are

    consistent with the observations from

    these case series. In one study, pregnant

    mice were given streptozotocin intra-

    peritoneally on gestational day 2 to

    induce diabetes, and the rate of severalcongenital anomalies (including HPE)

    among offspring was observed. Diabetic

    mice had an increased risk of delivering

    offspring with HPE; this risk was

    reduced among mice treated with insu-

    lin. To observe the interaction between

    diabetes and ethanol, the authors also

    exposed the pregnant diabetic mice to

    ethanol, and a decrease in the rate of

    HPE was observed. This led the authors

    to hypothesize that ethanol exposure

    might result in increased pregnancy

    loss of severely malformed embryos

    [Padmanabhan and Shafiullah, 2004].

    Four epidemiologic studies of pre-

    existing diabetes [Croen et al., 2000;

    Anderson et al., 2005; Orioli and

    Castilla, 2007; Correa et al., 2008] and

    three of four studies of gestational

    diabetes [Martinez-Frias et al., 1998;

    Croen et al., 2000; Anderson et al.,

    2005; Correa et al., 2008] have observed

    the prevalence of diabetes among moth-

    ers of infants with HPE to be at least

    twice as high as among control mothers

    (Table I). In the study using CBDMP

    data, the increased risk associated

    with ever having diabetes was mainly

    observed among mothers who reported

    taking insulin to treat their diabetes

    (odds ratio [OR] 10.2, 95% confidenceinterval [CI] 1.939.4 for mother taking

    insulin, OR 0.8, 95% CI 0.22.8 for

    those not taking insulin) [Croen et al.,

    2000]. A study using data from the

    NBDPS noted that the association

    between HPE and pregestational diabe-

    tes to be stronger when HPE occurred

    with other unrelated birth defects (OR

    16.2; 95% CI 1.6163.9), compared to

    when HPE occurred as an isolated defect

    (OR 6.0, 95% CI 0.749.8) [Correa

    et al., 2008]. However, CIs were wide

    because the number of cases with HPE

    was small.

    Despite the fact that diabetes is a

    well-recognized risk factor for HPE and

    other birth defects, its mechanism of

    action as a teratogen is not well under-

    stood. Hyperglycemia, oxidative stress,

    and other metabolic disturbances

    have been investigated as potential

    mechanisms for diabetic teratogenesis

    [Kousseff, 1999; Ornoy, 2007]. Better

    understanding of the mechanism

    could lead to improved therapeuticoptions for pregnant women with dia-

    betes to reduce their risk for birth

    defects; although reduced risks of

    birth defects appear to be achievable

    through better preconceptional care and

    glycemic control of diabetes during

    pregnancy, diabetic women continue

    to have an increased risk for having

    children with birth defects [Suhonen

    et al., 2000; Ray et al., 2001; Platt et al.,

    2002].

    Infections

    A variety of infections have beenreported

    among mothers of infants with HPE [see

    Cohen and Shiota, 2002 for a review].

    Congenital cytomegalovirus (CMV)

    infection has been described as a possible

    cause of HPE in several case reports andcase series [Barr et al., 1983; Byrne et al.,

    1987; Rasmussen et al., 1996; Kilic and

    Yazici, 2005]. However, documentation

    of congenital CMV was incomplete in

    several of these cases. Other infections

    noted in mothers who have had a child

    with HPEincludeone case where rubella

    was suspected in the mother, who had a

    rash and high fever during the first weeks

    of fetal development [Khudr and Olding,

    1973], and two infants with HPE and

    trisomy 13 who showed evidence of

    herpes simplex virus and of congenital

    syphilis [Rasmussen et al., 1996]. As with

    CMV, documentation of these congenital

    infections was incomplete.

    Three epidemiologic studies have

    investigated risk of HPE among mothers

    with respiratory illnesses, respiratory

    infections, or influenza during preg-

    nancy. While the study from ECLAMC

    identified an increased risk among

    mothers with influenza (OR 2.8, 95%

    CI 1.07.9) and an increased risk among

    mothers with respiratory illness wasreported using data from CBDMP (OR

    2.0, 95% CI 0.94.5) [Croen et al., 2000;

    Orioli and Castilla, 2007], a subsequent

    study from NBDPS found no association

    with respiratory infection or with fever

    during pregnancy [Miller et al., 2010].

    Studies using data from NBDPS

    and CBDMP have each suggested an

    increased risk of HPE among mothers

    with sexually transmitted infections.

    NBDPS investigators reported an OR

    of 2.1 (95% CI 0.95.1) for any sexuallytransmitted infection [Miller et al.,

    2010], while CBDMP researchers

    reported a prevalence of genital herpes

    or chlamydia of 4/48 among HPE case

    mothers and 2/107 among control

    mothers [Croen et al., 2000].

    High Blood Pressure

    Associations between HPE and mater-

    nal high blood pressure have been

    ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) 75

  • 7/29/2019 HPE Nongenetic Risk Factors

    4/13

    TABLEI.EpidemiologicStudiesofHoloprosencephalyandPre-ExistingandGestationalDiabetes

    Study

    Stud

    ydesign

    Sourceofcaseandexposureinformation

    Numberofcases

    OR(95%CI)

    Comments

    Pre-existingdiabetes

    Correaetal.[2008]

    Population-based

    case

    control

    study

    NationalBirthDefectsPreventionStudy,

    19972003,throughbirthdefects

    surveillancesystemsandmaternalinterview

    Isolated:42,

    multiple:19

    Isolated:6.00(0.7249.76),

    multiple:16.16

    (1.59163.88)

    Excluded

    recognized

    syndrom

    iccases,adjustedfor

    age,rac

    e/ethnicity,prenatal

    care,bo

    dymassindex,study

    center,householdincome

    OrioliandCastilla

    [2007]

    Hospital-based

    case

    control

    study

    LatinAmericanCollaborativeStudyon

    CongenitalMalformation

    s,19672000,

    throughhospital-basedre

    gistryandmaternal

    interview

    Allcases:281,

    isolated:182

    Allcases:3.5(0.916.2),

    isolated:1.9(0.49.7)

    Matchedonsex,location,

    andtim

    eofbirth,excluding

    chromo

    somalabnormalities

    Andersonetal.[2005]

    Population-based

    case

    control

    study

    TexasBirthDefectsMonito

    ringProgram,

    19972001,throughbirthdefects

    surveillanceandmaternalinterview

    57

    46.6(9.5230)

    Adjustedforethnicity,body

    massindex,age,education,

    smoking,alcoholuse,folic

    aciduse

    Croenetal.[2000]

    Population-based-

    based

    case

    contro

    lstudy

    CaliforniaBirthDefectsMonitoringProgram,

    19931996throughmed

    icalrecordreview

    andmaternalinterview

    48

    All:2.0(0.85.2),taking

    medication:10.2

    (1.939.4),no

    medication:0.8(0.22.8)

    Excluded

    caseswith

    chromo

    somalabnormalities

    Gestationaldiabetes

    Correaetal.[2008]

    Population-based

    case

    control

    study

    NationalBirthDefectsPreventionStudy,

    19972003,throughbirthdefects

    surveillancesystemsandmaternalinterview

    42

    0.76(0.105.75)

    Isolatedcases,excluded

    recognizedsyndromiccases,

    adjusted

    forage,race/

    ethnicit

    y,prenatalcare,body

    massindex,studycenter,

    householdincome

    Andersonetal.[2005]

    Population-based

    case

    control

    study

    TexasBirthDefectsMonito

    ringProgram,

    19972001,throughbirthdefects

    surveillanceandmaternalinterview

    47

    2.9(1.08.4)

    Adjustedforethnicity,body

    massindex,age,education,

    smoking,alcoholuse,folic

    aciduse

    Croenetal.[2000]

    Population-based

    case

    control

    study

    CaliforniaBirthDefectsMonitoringProgram,

    19931996throughmed

    icalrecordreview

    andmaternalinterview

    48

    2.0(0.75.3)

    Excluded

    caseswith

    chromo

    somalabnormalities

    Martinez-Friasetal.

    [1998]

    Hospital-based

    case

    control

    study

    LatinAmericanCollaborativeStudyon

    CongenitalMalformation

    s,19761995,

    throughhospital-basedre

    gistryandmaternal

    interview

    109

    3.27(1.297.77)

    Matchedonsex,location,

    andtim

    eofbirth

    OR,oddsratio;CI,confidenceinte

    rval.

    76 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) ARTICLE

  • 7/29/2019 HPE Nongenetic Risk Factors

    5/13

    investigated in epidemiologic studies

    using data from the NBDPS and

    CBDMP, neither of which identified

    an association [Croen et al., 2000; Miller

    et al., 2010].

    Anemia

    An increased risk of HPE to mothers

    with anemia was reported using data

    from the CBDMP (OR 3.4, 95% CI

    1.011.8) [Croen et al., 2000], although

    a subsequent study using data from

    ECLAMC found no such association

    (OR 0.9, 95% CI 0.15.9) [Orioli and

    Castilla, 2007] The ECLAMCstudyalso

    examined HPE and use of iron supple-

    ments during pregnancy, and no associ-

    ation was seen [Orioli and Castilla,

    2007].

    Other Maternal Illnesses

    and Injuries

    Researchers used CBDMP data to

    investigate associations between HPE

    and skin problems, allergies, urinary

    tract infection, and injury. No associa-

    tions were identified [Croen et al.,

    2000]. A single case of lobar HPE has

    been reported born to a mother with

    phenylketonuria who had high phenyl-

    alanine levels throughout the first 8

    10 weeks of pregnancy [Keller et al.,

    2000].

    MATERNAL THERAPEUTICEXPOSURES

    Salicylates, Including Aspirin

    Three case reports of children with HPE

    among mothers taking large doses of

    aspirin during early pregnancy have

    been described, and two additional cases

    have been described born to mothers

    taking sodium salicylate and sulfasalazine

    (Table II) [Khudr and Olding, 1973;

    Benawra et al., 1980; Agapitos et al.,

    1986; Koyama et al., 1996; Sezgin et al.,

    2002]. In these cases, the mothers had

    additional potential risk factors such as

    parental consanguinity [Sezgin et al.,2002], smoking [Benawra et al., 1980],

    and useof other medications [Khudr and

    Olding, 1973; Koyama et al., 1996]

    (Table II).

    Salicylate use has been investigated

    in three epidemiologic studies with

    inconsistent results. Findings from the

    CBDMP and NBDPS showed that

    mothers who have had a child with

    HPE were over twice as likely to have

    taken aspirin or salicylates during early

    pregnancy (Table III) [Croen et al.,

    2000; Miller et al., 2010], whereas the

    study from ECLAMC showed no asso-

    ciation [Orioli and Castilla, 2007].

    Findings from the CBDMP

    and NBDPS showed that

    mothers who have had a childwith HPE were over twice as

    likely to have taken aspirin or

    salicylates during early

    pregnancy (Table III), whereas

    the study from ECLAMC

    showed no association.

    As with all studies showing a relation

    between use of a medication and birthdefects, it is important to consider

    whether the observed association

    between salicylates and HPE observed

    in some studies is related to the medi-

    cation itself or to the indication for

    which the medication was used.

    Other Analgesics

    One case report has described HPE in an

    infant born to a mother who took

    acetaminophen in her second trimester

    of pregnancy, but the mother had also

    TABLE II. Case Reports of Holoprosencephaly Where Maternal Salicylate Use Was Reported

    Study Maternal characteristics Salicylate exposure Diagnosis

    Other reported exposures

    or risk factors

    Sezgin et al. [2002] 38-year-old with 2

    healthy children

    Aspirin 500g twice daily for

    1 week during first trimester

    Cyclopia, astomia, agnathia

    (46, XX)

    Consanguinity (parents first

    cousins)

    Agapitos et al. [1986] 41-year-old G2P1 Aspirin up to 4 g/day during

    first trimester

    Cyclopia, proboscis, multiple

    other birth defects, no

    chromosomal studies available

    Advanced age of mother (41)

    and father (56)

    Benawra et al. [1980] 20-year-old G2P0 Aspirin 34.5 g/day during firsttrimester Cyclopia, polypoid proboscis,multiple other birth defects,

    normal chromosomes

    Acetaminophen 610 tablets/day during second trimester,

    smoking 2 packs/day

    throughout pregnancy

    Khudr and

    Olding [1973]

    22-year-old G2P1 Sodium salicylate 3.6 g/day for

    7 days during first trimester

    Cyclopia, proboscis,

    low-set ears

    Viremia during week 3 of fetal

    development, treatment with

    high-dose penicillin G

    procaine, cortisone

    Koyama et al. [1996] 32-year-old G0P0 with

    Crohns disease

    Sulfasalazine 3 g/day before and

    during pregnancy

    Semilobar holoprosencephaly

    (46, XX)

    Ferrostatin (Fe) 100 mg/day,

    kernac (plant extract)

    240 mg/day, protective drug

    for gastritis or gastric ulcer,

    artificial insemination

    ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) 77

  • 7/29/2019 HPE Nongenetic Risk Factors

    6/13

    taken aspirin in the first trimester

    and smoked throughout pregnancy

    [Benawra et al., 1980]. Results from

    the NBDPS found a weak association

    between acetaminophen use during

    pregnancy and HPE (OR 1.5, 95% CI

    0.92.5), but no association with

    ibuprofen use [Miller et al., 2010].

    Antiepileptic Medications

    Kotzot et al. [1993] reported an infant

    with HPE whose mother was treated

    with phenytoin and primidone. In

    addition to HPE, the infant had features

    of fetal hydantoin syndrome, including

    hypertelorism, midfacial hypoplasia, and

    hypoplastic distal phalanges and nails of

    the fingers and toes [Kotzot et al., 1993].

    To address this possible association, a

    review of infants born at Boston area

    hospitals who were exposed to antiepi-

    leptic medications was performed, and

    one case of HPE was recorded among

    112 antiepileptic-exposed pregnancies.

    However, in a subsequent analysis of 453

    antiepileptic-exposed pregnancies, no

    cases with HPE were identified [Holmes

    and Harvey, 1994]. Rosa subsequently

    reviewed reports of adverse events to the

    U.S. Food and Drug Administration

    (FDA). He identified seven cases ofsuspected HPE including the case

    reported by Kotzot et al. [1993]

    among mothers taking antiepileptic

    drugs during pregnancy, including car-

    bamazepine, valproate, phenytoin, and

    primidone (used alone or in combina-

    tion) [Rosa, 1995]. The proportion of

    infants with HPE was higher among

    antiepileptic users than in mothers

    exposed to medications other than

    anticonvulsants.

    Retinoic Acid

    In a series of 154 mothers with prenatal

    exposure to isotretinoin, a retinoic

    acid derivative used to treat severe acne,

    21 had a child with a birth defect, one of

    whom had HPE [Lammer et al., 1985].

    Between 1969 and 1993, 25 cases of

    birth defects delivered following first-

    trimester exposure to topical tretinoin

    (all-trans-retinoic acid), a medication

    TABLEIII.EpidemiologicStudiesofHoloprosencephalyandMaternalSalicy

    lateUseinEarlyPregnancy

    Study

    Studydesign

    Sourceofcaseandexposure

    information

    Exposure

    Numberofcases

    OR(95%CI)

    C

    omments

    Milleretal.

    [2010]

    Population-based

    casecontrol

    study

    NBDPS,19972004,

    throughbirthdefects

    surveillanceandmaternal

    interview

    Aspirinduring

    periconceptionalperiod

    Allcases:72,

    isolated:50

    Allcases:3.4(1.66.9),

    isolated:2.0(0.75.5)

    Adjustedfor

    age,race/ethnicity,

    education,

    previous

    pregnancie

    s,folicacid,or

    multivitam

    inuse;excluded

    knownsyn

    dromiccasesand

    thosewhosemothershad

    pre-existingdiabetes

    Orioliand

    Castilla[2007]

    Hospital-based

    casecontrol

    study

    ECLAMC,19672000,

    throughhospital-based

    birthdefectsregistryand

    maternalinterview

    Aspirinorsalicylates

    Allcases:105,

    isolated:65

    Allcases:0.6(0.31.1),

    isolated:0.7(0.31.6)

    Matchedon

    sex,location,

    andtimeo

    fbirth,excluded

    caseswith

    chromosomal

    abnormalities

    Croenetal.[2000]Population-based

    casecontrol

    study

    CBDMP,19931996,

    throughmedicalrecord

    reviewandmaternal

    interview

    Aspirinorsalicylatesduring

    periconceptionalperiod

    48

    2.5(0.87.9)

    Excludedcaseswithchromosomal

    abnormalities

    OR,oddsratio;CI,confidenceinterval.

    78 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) ARTICLE

  • 7/29/2019 HPE Nongenetic Risk Factors

    7/13

    used for treatment of acne and of

    aging and sun-damaged skin, were

    reported to the FDA. Five of these

    25 had HPE, compared to 19 cases of

    HPE occurring in the 8,700 reports

    where the exposure was notretinoicacid

    [Rosa, 1994].

    Animal models also support a pos-

    sible relation between prenatal exposure

    to retinoic acid and HPE. Oral admin-

    istration of all-trans-retinoic acid in

    early pregnancy (gestational day 7) to

    inbred mice has produced a range of

    craniofacial phenotypes consistent with

    HPE [Kalter, 1992; Sulik et al., 1995;

    Lipinski et al., 2010].

    Antibiotics

    Case reports have described use ofantibiotics among mothers who have

    had a child with HPE, including

    penicillin procaine [Khudr and Olding,

    1973], ampicillin [Garzozi and Barkay,

    1985], and trimethoprimsulfamethox-

    azole [Ronen, 1992]. In the Kyoto

    embryo study, 0.8% of case mothers

    and 1.2% of control mothers reported

    using antibiotics during pregnancy

    [Matsunaga and Shiota, 1977].

    Hormones

    Case reports have described the use

    of several different hormones during

    early pregnancy among mothers who

    have had a child with HPE, including

    allyloestrenol and dihydrogesterone

    [Stabile et al., 1985], oral contraceptives

    [Ronen and Andrews, 1991], and

    cortisone [Khudr and Olding, 1973],

    although most motherstaking these hor-

    mones were taking other medications.

    Use of sex hormones was associated

    with an increased risk of HPE in thestudy using data from ECLAMC (OR

    3.7, 95% CI 1.311.0) [Orioli and

    Castilla, 2007]. An elevated risk was also

    seen in NBOPS among mothers using

    progesterone before or during the first

    trimester of programming, although the

    insert was not statistically significant

    (OR 2.3, 75%, CI 0.77.3) [Miller et

    al., 2010]. In the Kyoto embryo study,

    23% of case mothers had taken proges-

    tagens during early pregnancy compared

    to 15% of control mothers [Matsunaga

    and Shiota, 1977], while investigators

    from CBDMP found no association

    between HPE and use of birth control

    pills [Croen et al., 2000].

    Alkaloids Including Cyclopamineand Jervine

    In sheep and several other animal

    species, consumption of the corn lily

    Veratum californicum in early pregnancy

    was associated with malformations

    characteristic of HPE in the offspring

    [Coventry et al., 1998; Hovhannisyan

    et al., 2009; Welch et al., 2009].

    In sheep and several other

    animal species, consumption

    of the corn lily Veratum

    californicum in early pregnancy

    was associated with

    malformations characteristic of

    HPE in the offspring.

    Cyclopamine and jervine have since

    been identified as the primary terato-

    genic components of the corn lily plant.These alkaloids are structural analogs of

    cholesterol and act by inhibiting hedge-

    hog signaling [Hovhannisyan et al.,

    2009]. These components and related

    compounds are under investigation as

    therapeutic agents for treatment of

    psoriasis and several malignancies [Tas

    and Avci, 2004; Feldmann et al., 2008;

    Garber, 2008; Von Hoff et al., 2009]. We

    are not aware of any reports of human

    exposures to these alkaloids during

    pregnancy.

    Statins

    Several pieces of evidence have led to

    concerns that maternal exposure to

    statins, which are used to lower choles-

    terol levels, might raise the risk for HPE.

    These include the fact that some genetic

    conditions with defects in cholesterol

    metabolism include HPE as a feature in

    about 5% of cases [Kelley et al., 1996;

    Incardona and Roelink, 2000], inhib-

    ition of cholesterol biosynthesis has

    previously been shown to result in

    HPE in animal studies [Kolf-Clauw

    et al., 1997; Roux et al., 2000], and

    cholesterol is required for hedgehog

    signaling, a pathway which is critical in

    embryonic development and has

    been implicated in HPE [Roessler and

    Muenke, 2003; Edison and Muenke,

    2004b]. One report of 52 statin-exposed

    pregnancies reported to the FDA iden-

    tified 20 mothers who had a child with a

    major malformation, including 1 with

    HPE [Edison and Muenke, 2004a,b,

    2005]. However, among 22 mothers

    participating in the NBDPS and the

    Slone Epidemiology Center Birth

    Defects Study who were exposed to

    statins during pregnancy and had a child

    with a major birth defect, none had aninfant with HPE [Petersen et al., 2008].

    Analyses using a casecontrol design

    have not been possible because of the

    low frequency of exposure to statins

    during pregnancy.

    Other Medications

    Case reports and case series have

    described a wide variety of medication

    use during pregnancy among mothers

    who have had a child with HPE,

    including methotrexate (used for cancertreatment and as an abortifacient) [Seid-

    ahmed et al., 2006], tranexamic acid

    (used for the treatment of excessive

    bleeding) [Garzozi and Barkay, 1985],

    chlordiazepoxide (a benzodiazepine

    derivative used for the treatment of

    anxiety) [Ronen and Andrews, 1991],

    imipramine (a tricyclic antidepressant)

    [Ronen and Andrews, 1991], lithium

    (used in the treatment of bipolar

    disorder) [Rasmussen et al., 1996], and

    dimenhydrinate (an over-the-countermedication used to treat nausea and

    motion sickness) [Ronen, 1992]. A

    higher prevalence of medication use

    overall among case mothers compared

    to control mothers was observed in the

    Kyoto embryo study [Matsunaga and

    Shiota, 1977].

    Fertility Treatments

    Use of artificial insemination has been

    reported in a case of HPE, although the

    ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) 79

  • 7/29/2019 HPE Nongenetic Risk Factors

    8/13

    mother had other potential risk factors

    such as use of salicylates and other

    medications during pregnancy [Koyama

    et al., 1996]. Data from the NBDPS

    suggest an increased risk of HPE among

    mothers who have undergone assisted

    reproductive technologies (OR 4.2,

    95% CI 1.313.7), but not for fertility

    treatments overall (OR 0.9, 95% CI

    0.32.9) [Miller et al., 2010].

    Medical Exposures

    No association between HPE and

    X-rays or scans during pregnancy was

    observed by NBDPS or CBDMP inves-

    tigators [Croen et al., 2000; Miller et al.,

    2010]. Data from CBDMP were also

    used to investigate whether HPE was

    associated with surgery during preg-nancy, and no association was seen

    [Croen et al., 2000].

    MATERNALNON-THERAPEUTICEXPOSURES

    Alcohol

    Several case reports have implicated

    heavy alcohol consumption during the

    first trimester of pregnancy as a cause of

    HPE, although many of the reported

    mothers had other potential risk factors

    such as smoking and medication use

    [Bonnemann and Meinecke, 1990;

    Ronen and Andrews, 1991; Coulter

    et al., 1993; Rasmussen et al., 1996].

    Animal studies have supported a

    role for maternal alcohol exposure in the

    etiology of HPE. Acute maternal etha-

    nol exposure at an early stage (gastrula-

    tion) in embryonic development in mice

    has been shown to induce a spectrum of

    facial abnormalities consistent with HPE[Sulik and Johnston, 1982; Webster

    et al., 1983]. In a study of weekly

    Animal studies have supported

    a role for maternal alcohol

    exposure in the etiology of

    HPE. Acute maternal ethanol

    exposure at an early stage

    (gastrulation) in embryonic

    development in mice has been

    shown to induce a spectrum of

    facial abnormalities consistent

    with HPE.

    administration of ethanol throughout

    gestation to pregnant pigtail macaques,

    HPE was observed in a single fetus

    [Siebert et al., 1991]. Brief ethanol

    exposure of zebrafish during early gas-

    trulation has been shown to induce

    cyclopia [Blader and Strahle, 1998].

    Animal models have also been used to

    provide insight into the mechanism by

    which alcohol produces these abnor-

    malities in animals. These studies suggestthat ethanol exposure results in pertur-

    bation of expression of certain devel-

    opmental genes at a critical period in

    development [Blader and Strahle, 1998;

    Higashiyama et al., 2007].

    However, results of epidemiologic

    studies of maternal alcohol use during

    pregnancy have been inconsistent. An

    elevated risk (OR 2.0, 95% CI 0.94.5)

    was seen inthe study using CBDMP data

    [Croen et al., 2000], but researchers

    using data from NBDPS and the

    Kyoto embryo study found no associa-

    tion between alcohol use and HPE

    [Matsunaga and Shiota, 1977; Miller

    et al., 2010]. Using data from CBDMP,

    investigators noted an increased risk of

    HPE when mothers consumed alcohol

    and also smoked during pregnancy (OR

    5.4, 95% CI 1.420.0) [Croen et al.,

    2000].

    Smoking

    Several case reports have describedheavy smoking during pregnancy

    among mothers who had children with

    HPE, although most mothers had other

    potential risk factors such as alcohol

    intake and medication use [Benawra

    et al., 1980; Burck et al., 1982; Ronen

    and Andrews, 1991]. Three epidemio-

    logic studies have investigated the role of

    smoking as a potential r isk factor for

    HPE, with conflicting results. CBDMP

    researchers found a fourfold increased

    risk (OR 4.1, 95% CI 1.412.0)

    [Croen et al., 2000], whereas investiga-

    tors from the NBDPS and the Kyoto

    embryo study saw no association

    [Matsunaga and Shiota, 1977; Miller

    et al., 2010].

    Illicit Drugs

    Among mothers of infants with HPE

    identified through the New York State

    Congenital Malformations Registry,

    medical records indicated that 3 of 82

    used cocaine or marijuana during preg-

    nancy [Olsen et al., 1997]. Maternal

    marijuana use has been described in

    a case report of HPE, although this

    mother had also consumed alcohol

    heavily during pregnancy [Coulter

    et al., 1993]. No association between

    HPE and maternal illicit drug use has

    been seen in epidemiologic studies using

    data from NBPDS and CBDMP [Croen

    et al., 2000; Miller et al., 2010].

    NUTRITIONAL FACTORS

    Body Mass Index

    Three studies have investigated associa-

    tions between body mass index (BMI)

    and HPE. Increased risks were observed

    in studies using data from CBDMP (OR

    1.5, 95% CI 0.64.0 for BMI >29 kg/

    m2 vs. BMI 29 kg/m2) and from the

    Texas Birth Defects Registry (OR 1.4,

    95% CI 0.53.8 for BMI 30.0 kg/m2

    vs. BMI 18.524.9kg/m2, adjusted for

    ethnicity, age, education, smoking, alco-

    hol use, and folic acid use and excluding

    mothers with pre-existing diabetes),

    although CIs were wide and included

    the null value [Croen et al., 2000;

    Anderson et al., 2005]. The study from

    NBDPS, which also excluded women

    with diabetes from the analysis, found noassociation between maternal pre-preg-

    nancy obesity and HPE (OR 0.8, 95%

    CI 0.41.6 for BMI 30.0 kg/m2 vs.

    BMI 18.524.9 kg/m2, adjusted for

    age, race/ethnicity, education, previous

    pregnancies, folic acid use, and aspirin

    use) [Miller et al., 2010]. The

    Texas study also investigated interactions

    between gestational diabetes and obesity

    and observed that obesity alone

    (BMI 30.0kg/m2 compared to BMI

    80 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) ARTICLE

  • 7/29/2019 HPE Nongenetic Risk Factors

    9/13

  • 7/29/2019 HPE Nongenetic Risk Factors

    10/13

    [Matsunaga and Shiota, 1977; Croen

    et al., 2000].

    Gravidity

    Increased risk among mothers with high

    gravidity has been observed by CBDMP

    (OR 2.1, 95% CI 0.76.4 for gravidity 4vs. 1) and ECLAMC (OR 1.71, 95% CI

    0.943.10 for gravidity 4 or more vs.

    3 or less) investigators [Croen et al.,

    2000; Orioli and Castilla, 2007]. Based

    on data from NBDPS, no statistically

    significant association was observed

    between gravidity and HPE, although

    risk tended to increase with increasing

    gravidity [Miller et al., 2010].

    Previous Pregnancy Loss

    No association was observed betweenHPE and previous pregnancy loss in

    epidemiologic studies from CBDMP,

    ECLAMC, and NBDPS [Croen et al.,

    2000; Orioli and Castilla, 2007; Miller

    et al., 2010]. The Kyoto embryo study

    observed more miscarriages among case

    mothers compared to control mothers

    (29.5% vs. 21.0%), although the differ-

    ence was not statistically significant

    [Matsunaga and Shiota, 1977].

    Pregnancy Intention

    NBDPS investigators identified an

    increased risk for mothers whose pre-

    gnancies were unintended compared to

    those that were intended (OR 1.9, 95%

    CI 1.03.6), but not for mothers whose

    pregnancies were mistimed [Milleret al.,

    2010].

    Season of Conception

    Month of conception of HPE cases

    appeared to peak in March and has a

    low in September in the Kyoto embryostudy, although this finding was not

    statistically significant [Matsunaga and

    Shiota, 1977].

    SOCIODEMOGRAPHICFACTORS

    Age

    Maternal age has been inconsistently

    associated with HPE, with some studies

    finding lower risks among young

    mothers and higher risks among older

    mothers [Kallen et al., 1992; Orioli and

    Castilla, 2007], others finding increased

    risk among both younger and older

    mothers [Rasmussen et al., 1996; For-

    rester and Merz, 2000], and other studies

    finding no association with age [Matsu-

    naga and Shiota, 1977; Croen et al.,

    2000; Miller et al., 2010]. Some of this

    variation may be due to inclusion of

    infants with chromosome abnormalities

    (including trisomies) in some studies of

    HPE. No association between HPE and

    paternal age was seen in the three

    studies that investigated this association

    [Matsunaga and Shiota, 1977; Croen

    et al., 1996; Orioli and Castilla, 2007].

    Race/Ethnicity

    Higher risks of HPE among non-white

    compared to white mothers have been

    observed in four studies conducted in

    the United States and United Kingdom

    [Rasmussen et al., 1996; Olsen et al.,

    1997; Croen et al., 2000; Ong et al.,

    2007], while in a study in Hawaii, a

    greater risk among Far East Asians was

    seen, compared to Pacific Islanders (RR

    3.01, 95% CI 1.305.94) [Forrester and

    Merz, 2000]. No association was iden-

    tified between paternal race/ethnicity

    and HPE in the one study that investi-

    gated this association [Croen et al.,

    2000].

    CBDMP investigators observed an

    elevated risk among mothers born out-

    side the United States or Mexico (OR

    3.1, 95% CI 1.18.6) [Croen et al.,

    2000], while the analysis using data from

    the NBDPS found no increased risk to

    mothers born outside the United States

    [Miller et al., 2010].

    Socioeconomic Status

    Mothers with less than a high school

    education had a higher risk of HPE

    compared to mothers with a high school

    education by investigators using data

    from CBDMP (OR 2.1, 95% CI 0.8

    6.0) and from NBDPS (OR 2.5, 95% CI

    1.15.6) [Croen et al., 2000; Miller

    et al., 2010]. NBDPS investigators also

    noted a slightly increased risk for

    mothers with lower income (OR 1.5,

    95% CI 0.82.8), whereas the study

    using ECLAMC data found no asso-

    ciation between HPE and socioeco-

    nomic status [Orioli and Castilla, 2007;

    Miller et al., 2010].

    CONCLUSIONS

    Few non-genetic risk factors for HPE

    have been definitively identified to date,

    with diabetes being among the most

    commonly studied and the risk factor

    that has resulted in the strongest associ-

    ations, with fairly consistent results

    between studies. Other potential r isk

    Few non-genetic risk factors forHPE have been definitively

    identified to date, with diabetes

    being among the most

    commonly studied and the risk

    factor that has resulted in the

    strongest associations, with

    fairly consistent results

    between studies.

    factors that may warrant further study

    include salicylate use, respiratory and

    sexually transmitted infections, and use

    of assisted reproductive technologies,

    which have been identified in epide-

    miologic studies as possibly associated

    with increased r isk of HPE, and quality

    of diet and use of multivitamin supple-

    ments, which may be associated with a

    decreased risk.

    Caution is needed, however, in

    interpreting the results of epidemiologicstudies of HPE. Since the prevalence of

    HPE is much higher among embryos (4

    per 1,000) [Matsunaga and Shiota, 1977]

    than among live births (1 per 10,000)

    [Croen et al., 1996; Rasmussen et al.,

    1996], studying potential risk factors

    among live births or fetuses at 20 weeks

    or greater gestational age, as is com-

    monly done, may result in the identi-

    fication of exposures which are in fact

    not risk factors for HPE, but risk factors

    82 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) ARTICLE

  • 7/29/2019 HPE Nongenetic Risk Factors

    11/13

    for survival of the fetus long enough for

    detection of the anomaly.

    Small sample size leading to low

    power to detect an effect is another

    consequence of the low birth prevalence

    of HPE, which results in unstable effect

    estimates and wide CIs. These wide CIs,

    which frequently contain the null value

    (OR 1), make it difficult to identify

    which exposures may be true non-

    genetic risk factors for HPE as opposed

    to merely findings observed by chance.

    In this article some associations whose

    CIs include the null are referred to as

    potential risk factors when the effect

    estimate was fairly large or when more

    than one study found similar results;

    however, further research will be needed

    to determine whether these are true risk

    factors or not.Failure to exclude infants with

    conditions of genetic etiology (e.g.,

    chromosome abnormalities and single

    gene disorders) presents another

    challenge to studying non-genetic risk

    factors for HPE. Studies using etiolog-

    ically heterogeneous case groups may

    hamper identification of potential non-

    genetic r isk factors [Rasmussen et al.,

    2003]. For example, inclusion of differ-

    ent proportions of cases with chromo-

    somal anomalies may be responsible for

    the variability seen in studies of the

    association between maternal age and

    HPE. Similarly, given that maternal

    diabetes has been identified as a strong

    risk factor for HPE, studies aiming to

    identify other risk factors for HPE may

    benefit from exclusion of infants with

    diabetic mothers from the analysis.

    Another key factor in these studies

    is ensuring that all infants actually have

    HPE. Older studies often depended on

    facial phenotype for diagnosis of HPE,

    but some infants with HPE have normalfacies [Cohen, 1989a] and some infants

    with mild facial features consistent with

    HPE may not have HPE [Cohen,

    1989a,b]. Given the current widespread

    availability of neuroimaging, future

    studies should require that imaging

    studies (or pathologic analysis in

    deceased infants) documenting HPE be

    available. However, even when neuro-

    imaging is available, other brain defects

    may be mistaken for HPE: 19% of scans

    sent to the Carter Centers for Brain

    Research were found to not have

    HPE after review by radiologists with

    expertise in HPE [Stashinko et al.,

    2004].

    One consequence of a stricter case

    definition is further reduction in num-

    berof cases available for study. Analysis of

    the interaction between genetic and

    non-genetic risk factors will also be

    important, since some genotypes may

    increase the risk for HPE after certain

    exposures, but these studies will require

    even larger sample sizes. Future studies

    will likely require collaboration between

    multiple investigators to reach a sample

    size large enough to identify non-

    genetic risk factors for HPE.

    ACKNOWLEDGMENTS

    We would like to thank Dr. Jaime Fras

    for his helpful comments on an earlier

    draft of the manuscript.

    REFERENCES

    Agapitos M, Georgiou-Theodoropoulou M,Koutselinis A, Papacharalampus N. 1986.Cyclopia and maternal ingestion of salicy-lates. Pediatr Pathol 6:309310.

    Anderson JL, Waller DK, Canfield MA, ShawGM, Watkins ML, Werler MM. 2005.Maternal obesity, gestational diabetes, andcentral nervous system birth defects.Epidemiology 16:8792.

    Barr M Jr, Hanson JW, Currey K, Sharp S,Toriello H, Schmickel RD, Wilson GN.1983. Holoprosencephaly in infants ofdiabetic mothers. J Pediatr 102:565568.

    Becerra JE, Khoury MJ, Cordero JF, Erickson JD.1990. Diabetes mellitus during pregnancyand the risks for specific birth defects: Apopulation-based casecontrol study. Pedia-trics 85:19.

    Benawra R, Mangurten HH, Duffell DR. 1980.Cyclopia and other anomalies following

    maternal ingestion of salicylates. J Pediatr96:10691071.

    Bendavid C, Rochard L, Dubourg C, Seguin J,Gicquel I, Pasquier L, Vigneron J, Laquer-riere A, Marcorelles P, Jeanne-Pasquier C,Rouleau C, Jaillard S, Mosser J, Odent S,David V. 2009. Array-CGH analysis indi-cates a high prevalence of genomic rear-rangements in holoprosencephaly: Anupdated map of candidate loci. Hum Mutat30:11751182.

    Blader P, Strahle U. 1998. Ethanol impairs

    migration of the prechordal plate in thezebrafish embryo. Dev Biol 201:185201.

    Bonnemann C, Meinecke P. 1990. Holoprosen-cephaly as a possible embryonic alcoholeffect: Another observation. Am J MedGenet 37:431432.

    Bullen PJ, Rankin JM, Robson SC. 2001.Investigation of the epidemiology andprenatal diagnosis of holoprosencephaly inthe North of England. Am J Obstet Gynecol184:12561262.

    Burck U, Held KR, Kitschke HJ. 1982. Occur-rence of cyclopia, myelomeningocele, deaf-ness, and abducens paralysis in siblings. Am JMed Genet 11:443448.

    Byrne PJ, Silver MM, Gilbert JM, Cadera W,Tanswell AK. 1987. Cyclopia and congenitalcytomegalovirus infection. Am J Med Genet28:6165.

    Cohen MM Jr. 1989a. Perspectives on holopro-sencephaly: Part I. Epidemiology, genetics,

    and syndromology. Teratology 40:211235.

    Cohen MM Jr. 1989b. Perspectives on holopro-sencephaly: Part III. Spectra, distinctions,

    continuities, and discontinuities. Am J MedGenet 34:271288.

    Cohen MM Jr. 2006. Holoprosencephaly: Clin-ical, anatomic, and molecular dimensions.Birth Defects Res A Clin Mol Teratol76:658673.

    Cohen MM Jr, Shiota K. 2002. Teratogenesis of

    holoprosencephaly. Am J Med Genet 109:115.

    Correa A, Gilboa SM, Besser LM, Botto LD,Moore CA, Hobbs CA, Cleves MA,Riehle-Colarusso TJ, Waller DK, ReeceEA. 2008. Diabetes mellitus and birth

    defects. Am J Obstet Gynecol 199:237,e231e239.

    Coulter CL, Leech RW, Schaefer GB, ScheithauerBW, Brumback RA. 1993. Midlinecerebral dysgenesis, dysfunction of thehypothalamicpituitary axis, and fetal alco-hol effects. Arch Neurol 50:771775.

    Coventry S, Kapur RP, Siebert JR. 1998. Cyclop-amine-induced holoprosencephaly and asso-ciated craniofacial malformations in thegolden hamster: Anatomic and molecular

    events. Pediatr Dev Pathol 1:2941.Croen LA, Shaw GM, Lammer EJ. 1996.

    Holoprosencephaly: Epidemiologic andclinical characteristics of a Californiapopulation. Am J Med Genet 64:465472.

    Croen LA, Shaw GM, Lammer EJ. 2000. Riskfactors for cytogenetically normal holopro-sencephaly in California: A population-based casecontrol study. Am J Med Genet90:320325.

    Edison RJ, Muenke M. 2004a. Central nervoussystem and limb anomalies in case reports offirst-trimester statin exposure. N Engl J Med350:15791582.

    Edison RJ, Muenke M. 2004b. Mechanistic and

    epidemiologic considerations in the evalua-tion of adverse birth outcomes followinggestational exposure to statins. Am J MedGenet Part A 131A:287298.

    Edison RJ, Muenke M. 2005. Gestational expo-sure to lovastatin followed by cardiacmalformation misclassified as holoprosence-phaly. N Engl J Med 352:2759.

    Feldmann G, Fendrich V, McGovern K, Bedja D,Bisht S, Alvarez H, Koorstra JB, Habbe N,Karikari C, Mullendore M, Gabrielson KL,Sharma R, Matsui W, Maitra A. 2008. Anorally bioavailable small-molecule inhibitorof Hedgehog signaling inhibits tumorinitiation and metastasis in pancreatic cancer.Mol Cancer Ther 7:27252735.

    ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) 83

  • 7/29/2019 HPE Nongenetic Risk Factors

    12/13

    Forrester MB, Merz RD. 2000. Epidemiology ofholoprosencephaly in Hawaii, 198697.Paediatr Perinat Epidemiol 14:6163.

    Garber K. 2008. Hedgehog drugs begin to showresults. J Natl Cancer Inst 100:692697.

    Garzozi HJ, Barkay S. 1985. Case of true cyclopia.Br J Ophthalmol 69:307311.

    Higashiyama D, Saitsu H, Komada M, TakigawaT, Ishibashi M, Shiota K. 2007. Sequential

    developmental changes in holoprosence-phalic mouse embryos exposed to ethanolduring the gastrulation period. Birth DefectsRes A Clin Mol Teratol 79:513523.

    Holmes LB, Harvey EA. 1994. Holoprosence-phaly and the teratogenicity of anticonvul-sants. Teratology 49:82.

    Hovhannisyan A, Matz M, Gebhardt R. 2009.From teratogens to potential therapeutics:Natural inhibitors of the hedgehog signalingnetwork come of age. Planta Med. 75:

    13711380.Incardona JP, Roelink H. 2000. The role of

    cholesterol in Shh signaling and teratogen-induced holoprosencephaly. Cell Mol LifeSci 57:17091719.

    Kallen BA. 2005. Methodological issues in the

    epidemiological study of the teratogenicityof drugs. Congenit Anom (Kyoto) 45:4451.

    Kallen B, Castilla EE, Lancaster PA, MutchinickO, Knudsen LB, Martinez-Frias ML, Mas-troiacovo P, Robert E. 1992. The cyclopsand the mermaid: An epidemiological study

    of two types of rare malformation. J MedGenet 29:3035.

    Kalter H. 1992. Cyclopia produced in a very earlyretinoic acid experiment. Teratology 46:207208.

    Keller K, McCune H, Williams C, Muenke M.2000. Lobar holoprosencephaly in an infantborn to a mother with classic phenyl-ketonuria. Am J Med Genet 95:187188.

    Kelley RL, Roessler E, Hennekam RC, FeldmanGL, Kosaki K, Jones MC, Palumbos JC,Muenke M. 1996. Holoprosencephaly in

    RSH/Smith-Lemli-Opitz syndrome: Doesabnormal cholesterol metabolism affect thefunction of Sonic Hedgehog? Am J MedGenet 66:478484.

    Khudr G, Olding L. 1973. Cyclopia. Am J DisChild 125:120122.

    Kilic N, Yazici Z. 2005. A case of holoprosence-phaly and cebocephaly associated to torchinfection. Int J Pediatr Otorhinolaryngol69:12751278.

    Kolf-Clauw M, Chevy F, Siliart B, Wolf C,Mulliez N, Roux C. 1997. Cholesterolbiosynthesis inhibited by BM15.766 induces

    holoprosencephaly in the rat. Teratology 56:

    188200.Kotzot D, Weigl J, Huk W, Rott HD. 1993.

    Hydantoin syndrome with holoprosence-phaly: A possible rare teratogenic effect.Teratology 48:1519.

    Kousseff BG. 1999. Diabetic embryopathy. CurrOpin Pediatr 11:348352.

    Koyama N, Komori S, Bessho T, Koyama K,Hiraumi Y, Maeda Y. 1996. Holoprosence-phaly in a fetus with maternal medication ofsulfasalazine in early gestation. A case report.Clin Exp Obstet Gynecol 23:136140.

    Lammer EJ, Chen DT, Hoar RM, Agnish ND,Benke PJ, Braun JT, Curry CJ, FernhoffPM, Grix AWJr, Lott IT, et al. 1985.

    Retinoic acid embryopathy. N Engl J Med313:837841.

    Martinez-Frias ML, Bermejo E, Rodriguez-Pinilla E, Prieto L, Frias JL. 1998. Epide-miological analysis of outcomes of preg-nancy in gestational diabetic mothers. Am JMed Genet 78:140145.

    Matsunaga E, Shiota K. 1977. Holoprosencephalyin human embryos: Epidemiologic studies

    of 150 cases. Teratology 16:261272.Miller EA, Rasmussen SA, Siega-Riz AM, Frias

    JL, Honein MA, The National Birth DefectsPrevention Study. 2010. Risk factors fornon-syndromic holoprosencephaly in theNational Birth Defects Prevention Study.Am J Med Genet Part C Semin Med Genet154C:6272.

    Ming JE, Muenke M. 2002. Multiple hits duringearly embryonic development: Digenic dis-eases and holoprosencephaly. Am J Hum

    Genet 71:10171032.Muenke M, Gropman A. 2008. Holoprosence-

    phaly overview, GeneReviews, URL:http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?bookgene&parthpe-overview [accessedOctober 14, 2009].

    Munke M, Emanuel BS, Zackai EH. 1988.Holoprosencephaly: Association with inter-stitial deletion of 2p and review of thecytogenetic literature. Am J Med Genet30:929938.

    Odent S, Le Marec B, Munnich A, Le Merrer M,Bonaiti-Pellie C. 1998. Segregation analysis

    in nonsyndromic holoprosencephaly. Am JMed Genet 77:139143.

    Olsen CL, Hughes JP, Youngblood LG, Sharpe-Stimac M. 1997. Epidemiology of holopro-sencephaly and phenotypic characteristics ofaffected children: New York State, 19841989. Am J Med Genet 73:217226.

    Ong S, Tonks A, Woodward ER, Wyldes MP,Kilby MD. 2007. An epidemiological study

    of holoprosencephaly from a regional con-genital anomaly register: 19952004. PrenatDiagn 27:340347.

    Orioli IM, Castilla EE. 2007. Clinical epidemio-logic study of holoprosencephaly in SouthAmer ica. Am J Med Genet Par t A143A:30883099.

    Ornoy A. 2007. Embryonic oxidative stress as amechanism of teratogenesis with specialemphasis on diabetic embryopathy. ReprodToxicol 24:3141.

    Padmanabhan R, Shafiullah M. 2004. Effect ofmaternal diabetes and ethanol interactionson embryo development in the mouse. MolCell Biochem 261:4356.

    Petersen EE, Mitchell AA, Carey JC, Werler MM,

    Louik C, Rasmussen SA. 2008. Maternal

    exposure to statins and risk for birth defects:A case-series approach. Am J Med GenetPart A 146A:27012705.

    Platt MJ, Stanisstreet M, Casson IF, Howard CV,Walkinshaw S, Pennycook S, McKendrickO. 2002. St Vincents Declaration 10 yearson: Outcomes of diabetic pregnancies.Diabet Med 19:216220.

    Rasmussen SA, Moore CA, Khoury MJ, CorderoJF. 1996. Descriptive epidemiology of hol-oprosencephaly and arhinencephaly in met-ropolitan Atlanta, 19681992. Am J Med

    Genet 66:320333.Rasmussen SA, Olney RS, Holmes LB, Lin AE,

    Keppler-Noreuil KM, Moore CA. 2003.

    Guidelines for case classification for theNational Birth Defects Prevention Study.Birth Defects Res A Clin Mol Teratol67:193201.

    Rasmussen SA, Hayes EB, Jamieson DJ, OLearyDR. 2007. Emerging infections and preg-nancy: Assessing the impact on the embryo

    or fetus. Am J Med Genet Par t A143A:28962903.

    Ray JG, OBrien TE, Chan WS. 2001. Precon-ception care and the risk of congenitalanomalies in the offspring of women withdiabetes mellitus: A meta-analysis. Q J Med94:435444.

    Roach E, Demyer W, Conneally PM, Palmer C,Merritt AD. 1975. Holoprosencephaly:Birth data, genetic and demographic analy-ses of 30 families. Birth Defects Orig ArticSer 11:294313.

    Roessler E, Muenke M. 2003. How a Hedgehog

    might see holoprosencephaly. Hum MolGenet 12:R15R25.

    Roessler E, Belloni E, Gaudenz K, Jay P, Berta P,Scherer SW, Tsui LC, Muenke M. 1996.Mutations in the human Sonic Hedgehoggene cause holoprosencephaly. Nat Genet

    14:357360.Ronen GM. 1992. Holoprosencephaly and

    maternal low-calorie weight-reducing diet.Am J Med Genet 42:139.

    Ronen GM, Andrews WL. 1991. Holoprosence-phaly as a possible embryonic alcohol effect.Am J Med Genet 40:151154.

    Rosa F. 1994. Holoprosencephaly with first tri-mester topical tretinoin. Teratology 49:418.

    Rosa F. 1995. Holoprosencephaly and antiepilep-tic exposures. Teratology 51:230.

    Roux C, Wolf C, Mulliez N, Gaoua W, CormierV, Chevy F, Citadelle D. 2000. Role ofcholesterol in embryonic development. Am

    J Clin Nutr 71:1270S1279S.Scialli AR, Buelke-Sam JL, Chambers CD,

    Friedman JM, Kimmel CA, Polifka JE,Tassinari MS. 2004. Communicating risksduring pregnancy: A workshop on the use of

    data from animal developmental toxicitystudies in pregnancy labels for drugs. BirthDefects Res A Clin Mol Teratol 70:712.

    Seidahmed MZ, Shaheed MM, AbdulbasitOB, Al Dohami H, Babiker M, AbdullahMA, Abomelha AM. 2006. A case ofmethotrexate embryopathy with holopro-sencephaly, expanding the phenotype. BirthDefects Res A Clin Mol Teratol 76:138142.

    Sezgin I, Sungu S, Bekar E, Cetin M, Ceran H.2002. Cyclopiaastomiaagnathiaholo-

    prosencephaly association: A case report.

    Clin Dysmorphol 11:225226.Siebert JR, Astley SJ, Clarren SK. 1991. Hol-

    oprosencephaly in a fetal macaque (Macacanemestrina) following weekly exposure toethanol. Teratology 44:2936.

    Stabile M, Bianco A, Iannuzzi S, Buonocore MC,Ventruto V. 1985. A case of suspectedteratogenic holoprosencephaly. J Med Genet22:147149.

    Stashinko EE, Clegg NJ, Kammann HA, SweetVT, Delgado MR, Hahn JS, Levey EB.2004. A retrospective survey of perinatal risk

    factors of 104 living children with holopro-sencephaly. Am J Med Genet Part A 128A:114119.

    84 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) ARTICLE

  • 7/29/2019 HPE Nongenetic Risk Factors

    13/13

    Suhonen L, Hiilesmaa V, Teramo K. 2000.Glycaemic control during early pregnancyand fetal malformations in women with typeI diabetes mellitus. Diabetologia 43:7982.

    Sulik KK, Johnston MC. 1982. Embryonic originof holoprosencephaly: Interrelationship ofthe developing brain and face. Scan ElectronMicrosc, 1:309322.

    Sulik KK, Dehart DB, Rogers JM, Chernoff N.1995. Teratogenicity of low doses of all-transretinoic acid in presomite mouse embryos.Teratology 51:398403.

    Tas S, Avci O. 2004. Rapid clearance of psoriaticskin lesions induced by topical cyclopamine.

    A preliminary proof of concept study.Dermatology 209:126131.

    Von Hoff DD, LoRusso PM, Rudin CM, ReddyJC, Yauch RL, Tibes R, Weiss GJ, BoradMJ, Hann CL, Brahmer JR, Mackey HM,Lum BL, Darbonne WC, Marsters JC Jr, deSauvage FJ, Low JA. 2009. Inhibition of the

    hedgehog pathway in advanced basal-cellcarcinoma. N Engl J Med 361:1164

    1172.Webster WS, Walsh DA, McEwen SE, Lipson

    AH. 1983. Some teratogenic properties ofethanol and acetaldehyde in C57BL/6Jmice: Implications for the study of the fetalalcohol syndrome. Teratology 27:231243.

    Welch KD, Panter KE, Lee ST, Gardner DR,Stegelmeier BL, Cook D. 2009. Cyclop-amine-induced synophthalmia in sheep:Defining a critical window and toxicoki-netic evaluation. J Appl Toxicol 29:414421.

    Whiteford ML, Tolmie JL. 1996. Holoprosence-

    phaly in the west of Scotland 19751994. JMed Genet 33:578584.

    Yoon PW, Rasmussen SA, Lynberg MC, MooreCA, Anderka M, Carmichael SL, Costa P,Druschel C, Hobbs CA, Romitti PA,Langlois PH, Edmonds LD. 2001. TheNational Birth Defects Prevention Study.Public Health Rep 116:3240.

    ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS) 85