IN UTERO EXPOSURE TO VALPROATE INCREASES THE RISK FOR ISOLATED CLEFT PALATE
Adam Jackson1, Rebecca Bromley2, James Morrow3, Beth Irwin3, Jill Clayton-Smith4
1Blackpool Victoria Hospital, Whinney Heys Road, Blackpool, Lancashire, United Kingdom, FY3 8NR
2The University of Manchester, Institute of Human Development, United Kingdom, M13 9PL
3Neurology Department, Belfast Health and Social Care Trust, Royal Victoria Hospital, Belfast, Co Antrim, Northern Ireland, BT12 6BA
4Manchester Centre For Genomic Medicine, Central Manchester University Hospitals, Manchester, United Kingdom, M13 9WL
Correspondence to:
Adam Jackson
Postal address: Blackpool Victoria Hospital, Whinney Heys Road, Blackpool, UK, FY3 8NR
Email:
Telephone: 07976056436
Keywords:
Valproate, orofacial, clefting, palate, pregnancy
Word Count (excluding title page, abstract, references, figures and tables): 2925
ABSTRACT:
INTRODUCTION: Orofacial clefting (OFC) has been described in infants exposed to sodium valproate (VPA) prenatally, but often no distinction is made between cleft lip and palate (CLP), and cleft palate only (CPO). This distinction is important as these conditions have different management implications and the distinction has implications too for understanding the teratogenic mechanisms .
METHODS: Using the Pubmed database, case reports and observational studies describing OFC in association with VPA exposure were identified. Searches for similarly exposed patients referred to a regional genetic centre and those recorded in the UK Epilepsy and Pregnancy Register (UKEPR) were undertaken. Cleft type and, where available, VPA doses prescribed were recorded.
RESULTS: Twelve out of fifteen published cases had CPO and five out of ten observational studies showed an increase in CPO compared to CLP. Eight patients ascertained through the regional genetic centre had CPO. Thirteen cases of OFC occurred in 1282 VPA monotherapy exposed pregnancies on the UKEPR; nine had CPO and four had CLP, representing an 11.3-fold and 3.5-fold increase risk in CPO and CLP, respectively over general population risk. Doses ranged from 200-2500mg VPA daily with 73% of monotherapy CPO cases from the local cohort and UKEPR occurring at doses over 1000mg.
CONCLUSIONS: CPO is the predominant cleft-type seen in prenatal VPA exposure. Parents should be counselled appropriately and infants should undergo review after delivery for CPO. Pregnancy registers collecting information on congenital anomalies should make the distinction between CLP and CPO as the risk differs across the two conditions.
INTRODUCTION
Foetal Valproate Syndrome (FVS) describes a constellation of features that has been recognised in a number of children exposed to valproic acid (VPA) in utero. First delineated in 1984 by DiLiberti et al.[1], FVS is characterised by a pattern of major and minor malformations and neurodevelopmental problems together with a distinctive facial phenotype. The latter includes medial deficiency of the eyebrows which tend to be neat and arched, a prominent infra-orbital groove, a broad and flat nasal bridge, a smooth philtrum with a thin vermillion border to the upper lip and a small, downturned mouth with an everted lower lip. Congenital abnormalities which occur at greater frequency in FVS include neural tube defects, radial ray defects, trigonocephaly and heart defects. A dose dependent effect of VPA has been established,[2] with doses above 1000mg daily being associated with a greater risk of congenital malformations. The risk of cognitive impairment is increased with increasing dose.[3]
Orofacial clefting (OFC) has been associated with exposure to VPA in utero[4] and has been described as a feature of FVS.[5] However, the spectrum of OFC includes both cleft lip and/or palate (CLP) and cleft palate only (CPO). The causes and developmental pathogeneses of these two conditions are different as are the management interventions required and outcomes expected after surgical correction. Children who have CLP are managed on a different care pathway from those who have CPO. The number and types of surgical interventions will differ, for example, and children with an isolated cleft lip are not likely to need the same speech and language interventions or to have the middle ear complications that are so frequent in patients with cleft palate. The implications for the two conditions are thus different and to provide correct information to parents about cause, management, complications and prognosis it is important to make the distinction between CLP and CPO. In the case of FVS, many patient information leaflets, websites, published papers and textbooks link VPA exposure in pregnancy with an increased incidence of OFC but do not go into details as to whether this association is more likely to be with CLP or CPO.[6],[7],[8]
From a teratological point of view it is also important to distinguish between the two types of OFC as there are different causal mechanisms[9]. Further, current practice in most malformation registries is to report a single incidence figure for OFC, combining both CLP and CPO; however if there are mechanistic differences the risks associated with each time of OFC may differ and combining them may lead to unreliable risk estimates.
In the 4th week of gestation, the first pair of pharyngeal arches gives rises to paired maxillary and mandibular prominences which approximate a single frontonasal prominence. In the next two weeks, the maxillary prominences grow medially and fuse with the medial nasal prominences forming the philtrum, upper jaw and the triangular primary palate. In the sixth week of development, the palatine shelves which are outgrowths of the maxillary prominences progress obliquely downward on each side of the tongue. It is in the seventh week that these shelves turn horizontally and fuse creating the secondary palate with the incisive foramen marking the division of the secondary and primary palates.[10]
Clefts anterior to the incisive foramen cause CLP, arising from incomplete fusion of the maxillary prominence with the medial nasal prominence. Clefts posterior to the incisive foramen cause CPO, arising from incomplete fusion of the palatine shelves. CPO can lead to velopharyngeal insufficiency and carries significant risk of chronic otitis media, hearing loss and abnormal speech, unlike CLP.[11]
Antenatal ultrasound scans (USS) are helpful for identifying congenital malformations and allow both parents and clinicians to make preparations for the management of these conditions after birth. In the UK every pregnant mother is offered a detailed fetal anomaly scan at around 20 weeks of pregnancy. Offerdal et al. (2008) studied 49 314 pregnancies in Norway and found a pickup rate of 43% (35 out of 77 cases) for CLP and 0% (0 out of 24 cases) for CPO.[12] A novel 3D sonographic technique proposed by Campbell et al. allowed for improved antenatal detection for CPO.[13] Overall, however, in a routine ante-natal setting a scan which is negative for CLP does not exclude CPO.
Martinez-Frias et al. 1990[14] stated that OFC only occurred in children exposed to VPA in utero as polytherapy. However, two large epidemiological studies (Jentink et al. 2010, Gilboa et al. 2011) have since found an increased odds ratio for CPO for children of mothers taking VPA monotherapy. The authors’ anecdotal experience of children exposed to VPA in utero who have attended both a cleft clinic and the regional genetic clinic in Manchester, together with the results of cohort studies of children exposed to antiepileptic drugs in utero carried out within the North West of England have suggested that CLP is, in fact, unusual after VPA exposure in utero and that CPO is in fact much more likely. Such observations led to this review of the available literature and of other known patients with OFC following VPA exposure.
METHODS
A literature search was undertaken to identify any case reports, case series or cohort studies that described OFC occurring in association with FVS. Searches were conducted in PubMed, from conception to March 2014 with no language restrictions using the search terms ‘fetal valproate syndrome’, ‘valproate cleft lip’, ‘valproate cleft palate’ and ‘anti-epileptic drugs pregnancy cleft lip palate’. Studies were filtered to include only those that described OFC occurring in VPA exposed patients, whether available in abstract or full-text form. A manual literature search was also undertaken to identify any additional reports and meeting abstracts in the investigators’ personal collections. All reports were reviewed and, where possible, information on VPA dosage during gestation was recorded. Finally, additional cases known to the authors through attendance at a regional genetic clinic were reviewed . Finally, data was obtained from the UK Epilepsy and Pregnancy Register about cases with a history of VPA exposure and OFC. All patient data was anonymised for analysis but cleft status remained linked to AED type and dose.
RESULTS
Cases reported in the literature
Report / No affected individuals / total reported / Doses / Type of OFCYaqoob et al.[15] / 1/1 / - / CPO
Schorry et al.[16] / 1/5 / 750mg VPA / CPO
Ozkinay et al.[17] / 1/1 / 1800mg VPA
100mg LTG / CPO
(XXX karyotype)
Zaki et al.[18] / 1/1 / 1200mg VPA / CPO
Boyle et al.[19] / 1/1 / 2300mg VPA / CPO
Epinasse et al.[20] / 1/4 / - / ‘Oral clefting’
Okada et al.[21] / 1/1 / - / CPO
Winter et al.[22] / 1/4 / 1500mg VPA
600mg CBZ / CPO
Halder et al.[23] / 1/1 / 900mg VPA / CPO
Malm et al.[24] / 2/7 / 2500mg VPA / CPO
2500mg VPA / CPO
Lindhout et al.[25] / 2/18 (cohorts A and B) / 600mg VPA, 800mg CBZ, 150mg PHB, 300mg PHT / CL
900mg VPA, 600mg CBZ, 150mg PHB, 300mg PHT / CLP
Eroglu et al.[26] / 1/2 / 1000mg VPA / CPO
Jackson et al.[27] / 1/4 / 1500mg VPA / CPO
Table 1: Case reports describing OFC in FVS. VPA, sodium valproate. LTG, lamotrigine. CBZ, carbemazepine. PHB, phenobarbitone. PHT, phenytoin.
32 publications describing children exposed to VPA in pregnancy were identified, describing 116 children. Thirteen papers reported the occurrence of OFC with 15 affected individuals (Table 1): 12 had CPO, one CL, one CLP and one reported as ‘oral clefting’ with no further information. Doses during pregnancy were available for 12/15 individuals, ranging from 600mg to 2500mg daily (mean = 1454mg daily). Polytherapy had been used in 4/12 cases (with lamotrigine, carbamazepine, phenobarbitone and phenytoin)
For those where drug dose was reported, 8/10 (80%) of CPO cases occurred in patients exposed to over 1000mg VPA in utero, whereas 2/2 (100%) of CLP cases occurred in patients exposed to less than 1000mg in utero. One of the affected CPO patients was reported to have a XXX karyotype, a condition which is not known to be associated specifically with CPO.
Study Reference / Population / OFC / CPO / CLPJentink et al.[28] – meta-analysis / 1565 VPA monotherapy / 13 / 13 / 0
Jentink et al. –
EUROCAT case-control study / 2244 VPA monotherapy / 13 / 13 (Odds Ratio, OR 5.2) / 0
Gilboa et al.[29] – meta-analysis / Meta-analysis of articles quoting association of VPA with birth defects
11 included / - / OR 5.8 / Not stated
Vajda et al.[30] / 447 VPA exposed (mono or polytherapy) / 5 / No distinction / No distinction
Endo et al.[31] / 25 VPA polytherapy / 1 / 0 / 1*
Moore et al.[32] / 34 VPA monotherapy / 0 / 0 / 0
Kozma et al.[33] – meta-analysis** / 70 VPA monotherapy / 3 / 3 / 0
Samren et al.[34] / 184 VPA monotherapy / 0 / 0 / 0
118 VPA polytherapy / 1 / No distinction / No distinction
Hernandez-Diaz et al.4 *** / 323 VPA monotherapy / 3 / 2 / 1
NOTE:
*In this case, the mother was taking 1400mg VPA, 1.5mg clonazepam and 200mg zonisamide daily.
**2 of these cases (Winter et al and Okada et al.) were reported above as case reports
*** Found in the supplementary outcome data
Table 2: Observational studies and Systematic Reviews describing OFC in FVS.
Seven observational studies and three literature reviews were identified using the search strategy described (Table 2). Jentink et al. (2010) reviewed the literature on malformations in FVS and included eight studies in a meta-analysis of 1565 patients. Three of these eight studies were also identified by the search strategy but have only been included as a composite of Jentink et al. The 14 most common malformations found in this meta-analysis were then assessed for association with VPA monotherapy exposure in the first trimester by a case-control study using the European Surveillance of Congenital Abnormalities (EUROCAT) database. This study found CPO to be significantly associated with VPA exposure in the meta-analysis and found that this association remained in the case-control study. Gilboa et al.( 2011) undertook another meta-analysis which found an increased odds ratio for CPO in VPA exposure.
Cleft-type was not distinguished by Vajda et al. (2013) or Samren et al. (1997). CPO was identified in 3/70 patients by Kozma et al. (2001) (general population risk 1 in 700, although there is regional variation[35]) and in 2/323 by Hernandez-Diaz et al. (2012). One case of CLP was noted by both Hernadez-Diaz et al (2012) and Endo et al. (2004). Moore et al. (2000) reported no patients with CLP or CPO when reviewing 34 cases of VPA exposure.
Few observational studies reported the dose taken by mothers during pregnancy. Only Endo et al. (2004) provided this information and in this case there was one CLP in a pregnancy exposed to 1400mg VPA daily.
In summary, 54 children with clefting following VPA exposure were identified from published cases with 43 of these cases being CPO.
Cases ascertained through the Regional Genetic Service