Prenatal testing for uniparental disomy: indications and clinical relevance


  • D. Kotzot

    Corresponding author
    1. Division of Clinical Genetics, Department of Medical Genetics, Molecular and Clinical Pharmacology, Medical University of Innsbruck, Innsbruck, Austria
    • Division of Clinical Genetics, Department of Medical Genetics, Molecular and Clinical Pharmacology, Schoepfstr. 41, A-6020 Innsbruck, Austria
    Search for more papers by this author


This review aims to provide a rational and ethical basis for prenatal testing for uniparental disomy (UPD) in cases with abnormal ultrasound findings or numeric and/or structural chromosomal aberrations in chorionic villous or amniotic fluid samples. The clinical phenotypes of the genomic imprinting-associated paternal UPD 6 (transient neonatal diabetes mellitus), maternal UPD 7 (Silver–Russell syndrome), paternal UPD 11p (Beckwith–Wiedemann syndrome), maternal UPD 14 (precocious puberty, short stature and highly variable developmental delay), paternal UPD 14 (polyhydramnios and a bell-shaped thorax), maternal UPD 15 (Prader–Willi syndrome), paternal UPD 15 (Angelman syndrome), maternal UPD 16 and UPD 20, as well as the diagnostic options, are summarized. In addition, the clinical impact of UPD testing and its relevance in various prenatal diagnostic situations are discussed. As a general rule, prenatal UPD testing, following genetic counseling, is justified if paternal UPD 14, maternal UPD 15 or paternal UPD 15 are suspected. In contrast, considering the mild phenotypes of paternal UPD 6 and maternal UPD 7, prenatal UPD testing is questionable. Because of the highly variable phenotype for paternal UPD 11p, maternal UPD 14 and maternal UPD 16, prenatal testing should be discussed critically on an individual basis. For all other chromosomes, prenatal UPD testing is purely academic and should therefore not be performed on a routine basis, particularly because a positive result might confuse the parents more than it actually helps them. Copyright © 2007 ISUOG. Published by John Wiley & Sons, Ltd.


The concept of uniparental disomy (UPD), the inheritance of a pair of chromosomes from one parent with no contribution from the other parent, was introduced into medical genetics by Engel1 in 1980. During the last two decades, the clinical impact of UPD and associated imprinting disorders, such as Prader–Willi syndrome, Angelman syndrome, Silver–Russell syndrome, Beckwith–Wiedemann syndrome and transient neonatal diabetes mellitus, increasingly have come to our attention2. It has become obvious that UPD per se is without clinical relevance; problems arise instead from the association of UPD with trisomy mosaicism, abnormal genomic imprinting or homozygosity of autosomal recessively inherited mutations, and the type of problem is in part dependent on the mechanism of formation of the UPD (trisomy rescue: the fertilization of a disomic gamete by a normal gamete followed by loss of the chromosome from the normal gamete; gamete complementation: the fertilization of a disomic gamete by a gamete nullisomic for the same chromosome; postfertilization error: mitotic nondisjunction and loss of the monosomic chromosome or vice versa; monosomy rescue: the fertilization of a normal gamete by a nullisomic gamete and somatic reduplication of the monosomic chromosome)3. Homozygosity of an autosomal recessively inherited mutation can be caused by isodisomy, defined as the presence of two copies of one parental chromosome. Isodisomy of the whole chromosome is obligatory in monosomy rescue and in postzygotic formation, both of which are rare in maternal UPD4. Isodisomy of a small segment of chromosome due to double recombination cannot be excluded in cases of heterodisomy, which is defined as the presence of both of the contributing parent's alleles. Genomic imprinting is used to describe the situation when expression of an autosomal gene occurs from only one allele; which of the two alleles is expressed is dependent upon its parental origin. It is estimated that around 100 genes are imprinted, although because they are almost always clustered, the number of imprinted chromosomal regions is low. So far, chromosomal loci 6q23, 7p15, 7q22, 7q32, 11p13, 11p15, 14q32 and 15q12 have been defined as imprinted by the associated clinical phenotypes and synteny with mice (Table 1), while for chromosomes 16 and 20q the data are inconsistent.

Table 1. Phenotype and etiology of uniparental disomy (UPD)-associated imprinting disorders
DisorderEtiology (% of positive cases)Phenotype
Transient neonatal diabetes mellitusPaternal UPD 6 (c. 50%)Transient neonatal diabetes mellitus, intrauterine growth restriction, macroglossia
 Imprinting defects on 6q24 (c. 10–20%) 
Silver–Russell syndromeMaternal UPD 7 (c. 5%)Pre- and postnatal growth restriction, retarded bone age, hemihypotrophy, facial dysmorphisms including relative macrocephaly, a triangular face with a high and broad forehead and a pointed chin
 Imprinting defects on 11p (c. 30%) 
 Structural chromosomal aberrations (c. 1%) 
Beckwith–Wiedemann syndromeImprinting defects on 11p (c. 60%)Pre- and postnatal overgrowth, abdominal wall defects, hypoglycemia, facial dysmorphisms including macroglossia and earlobe creases and posterior helical pits, risk of embryonic tumors
 Paternal UPD 11p (c. 20%) 
 Gene mutations (CDKN1C) (c. 10%) 
 Paternal duplications, inversions, translocations (c. 1–2%) 
Maternal UPD 14Maternal UPD 14 (c. 100%)Short stature, muscular hypotonia, precocious puberty, truncal obesity, variable psychomotor retardation
Paternal UPD 14Paternal UPD 14 (c. 100%)Severe psychomotor retardation, polyhydramnios, mild contractures of the fingers, bell-shaped thorax with a characteristic rib configuration (‘coat-hanger sign’)
Prader–Willi syndromeMicrodeletion (c. 70%)Muscular hypotonia, feeding difficulties in infancy followed by hyperphagia and subsequent obesity, moderate mental retardation, hypogonadotropic hypogonadism, facial dysmorphisms including almond-shaped eyes, short hands and feet
 Maternal UPD 15 (c. 25%) 
 Imprinting defect (c. 3%) 
Angelman syndromeMicrodeletion (c. 70%)Severe mental retardation, ataxia, seizures, electroencephalographic abnormalities, jerky movements, inappropriate laughter
 Paternal UPD 15 (c. 1–3%) 
 UBE3A mutations (c. 5–10%) 
 Imprinting defects (c. 3–5%) 

UPD itself along with the associated findings pose a great challenge, particularly for genetic counseling as well as for the clinical geneticist: which test should be performed in which clinical situation? To use all available tests and to overload the parents with information and facts is one possibility, but certainly not the best. In order to provide a basis for adequate prenatal management of such cases, clinical phenotypes and diagnostic options are summarized briefly here, and the clinical impact of UPD testing and its relevance in various prenatal situations are discussed.


The phenotype of paternal UPD 6 is characterized by transient neonatal diabetes mellitus, intrauterine growth restriction (IUGR) and macroglossia5. So far, almost 20 cases have been reported, most of them with a normal chromosomal complement4.

Maternal UPD 7 is found in approximately 5% of cases with a phenotype of Silver–Russell syndrome6 and is characterized by pre- and postnatal growth restriction, retarded bone age, hemihypotrophy and facial dysmorphisms including relative macrocephaly and a triangular face7, 8. Mild developmental delay has been described in around half of all cases9. So far, around 50 cases have been reported, most of them with a normal chromosomal complement4.

Segmental paternal UPD 11p due to postzygotic recombinations is found in around 20% of cases with Beckwith–Wiedemann syndrome10. The phenotype of Beckwith–Wiedemann syndrome is characterized by pre- and postnatal overgrowth, abdominal wall defects, hypoglycemia, facial dysmorphisms including macroglossia and earlobe creases/posterior helical pits, and an increased risk for embryonic tumors11.

The phenotype of maternal UPD 14 includes short stature, precocious puberty, postpubertal truncal obesity, small hands and feet and mild relative macrocephaly12. Of all types of UPD, psychomotor development is the most variable in UPD 14: cases with severe mental retardation as well as patients with normal psychomotor development have been documented. So far, more than 50 cases have been reported, around half of them with a homologous or a non-homologous Robertsonian translocation4.

Paternal UPD 14 shows the most severe UPD phenotype. Most cases die shortly after birth or in infancy. Polyhydramnios, IUGR, cardiomyopathy and a specific configuration of the thoracal ribs (‘coat hanger sign’) are characteristic features13. So far, almost 20 cases have been reported, many of them associated with a homologous or a non-homologous Robertsonian translocation4.

Maternal and paternal UPD 15 represent the most frequently observed UPDs. The phenotype of maternal UPD 15, which is found in around 25–30% of cases with Prader–Willi syndrome, is characterized by muscular hypotonia, feeding difficulties in infancy followed by hyperphagia with subsequent obesity, moderate mental retardation, hypogonadotropic hypogonadism and facial dysmorphisms14. Patients with paternal UPD 15 suffer from Angelman syndrome, which is characterized by severe mental retardation, ataxia, seizures, electroencephalographic abnormalities, jerky movements, inappropriate laughter and facial dysmorphisms, including progenia and a wide mouth15.

For maternal UPD 16 the data are inconsistent. Some patients are affected by IUGR with or without catch-up growth. A few cases have malformations such as cardiac defects or imperforate anus. So far, around 50 cases have been reported, many of them with prenatally diagnosed trisomy mosaicism4.

For maternal and paternal UPD 20, only four and one case, respectively, have been described4. All were associated with a numeric and/or structural chromosomal aberration. None of these cases had the phenotype of the maternally expressed GNAS gene, which causes pseudohypoparathyroidism type IB, or of the paternally expressed Neuronatin gene2, both of which are located on the long arm of chromosome 20. Therefore, the phenotypic consequences of UPD 20 are unknown at present.

As methylation is the major mechanism regulating genomic imprinting and gene expression, methylation-specific assays are the method of choice for investigation. However, due to etiological heterogeneity, these assays are not equally informative. For Prader–Willi syndrome, Angelman syndrome and UPD 14, almost all cases can be detected using methylation-specific assays. For Silver–Russell syndrome and Beckwith–Wiedemann syndrome, the detection rates are only c. 5% and c. 20%, respectively6, 10. The detection rate for transient neonatal diabetes mellitus is unknown. Furthermore, to distinguish UPD from microdeletions or imprinting defects in Prader–Willi syndrome and Angelman syndrome, an additional test is needed, usually microsatellite-based haplotype analysis and/or fluorescence in-situ hybridization (FISH) analysis with locus-specific probes. For chromosomes not carrying imprinted genes, microsatellite-mediated haplotype analysis is the method of choice. Generally, according to the recommendations of the American Society of Human Genetics, two informative markers per chromosome are accepted for diagnosis of both hetero- and isodisomy16.


To define indications for UPD testing in postnatal cases is usually not difficult; only the presence of a clinical phenotype as described in Table 1 should initiate UPD testing. Principally, the major reasons for testing are: (a) to make a diagnosis, (b) to offer better therapy to the patient, and (c) to provide more information to the parents regarding prognosis and recurrence risk. Therefore, if the only clinical manifestation is multiple miscarriages, as in the case of an exclusively heterochromatic supernumerary marker chromosome (SMC) other than inv dup(15) or in the case of a balanced non-Robertsonian translocation in one parent, UPD testing is of academic interest only, because UPD has no relevance for meiosis and there is no risk of recurrence in offspring.

The situation is more complex prenatally because of the possibility of pregnancy termination and the limited clinical information available. Two major situations can indicate UPD testing: (1) sonographic detection of growth anomalies and/or malformations resembling a UPD phenotype with no family history or other risk factors; (2) a known familial chromosomal aberration or unexpected findings on routine chorionic villus sampling (CVS) or amniocentesis. For all people involved the situation is unexpected and psychologically exceptional, and decisions must be made under the pressure of time.

Ultrasound findings

If life-threatening malformations or anomalies and/or IUGR are observed on ultrasound, the question of whether UPD is present is of little consequence to the decision that must be made. In contrast, sonographic findings such as asymmetrical growth restriction (resembling Silver–Russell syndrome), overgrowth in association with omphalocele (being features of Beckwith–Wiedemann syndrome), or a small thorax and polyhydramnios (typically found in paternal UPD 14) are not initially life-threatening and thus require a different diagnostic procedure from that required in the case of life-threatening malformations. Usually, CVS or amniocentesis is indicated to search for numeric and/or structural chromosomal aberrations. In the case of asymmetrical growth restriction and a normal karyotype, testing for maternal UPD 7 and an 11p imprinting defect17 should be discussed carefully. Both investigations together detect only c. 60% of cases with Silver–Russell syndrome and the prognosis for psychomotor development and final height is good under the conditions of birth surveillance in a central hospital (which in any case would be recommended because of the IUGR) and growth hormone therapy during childhood. In the case of omphalocele, overgrowth and macroglossia, the differential diagnosis of Beckwith–Wiedemann syndrome might be considered. However, most cases of Beckwith–Wiedemann syndrome develop well, both physically and mentally, although c. 20% die during the perinatal period, and some children with Beckwith–Wiedemann syndrome will face relevant medical problems11. Therefore, a total risk of c. 30% for an adverse outcome might justify prenatal diagnosis—not to indicate termination but to enable better management after delivery, because, in contrast to Silver–Russell syndrome, hypoglycemia and sometimes early manifesting tumors require adequate treatment shortly after birth. A small thorax with ribs resembling coat-hangers is a strong hint towards paternal UPD 14 and, because of the poor prognosis, should be tested in order to offer the choice of termination to the parents.

Numeric chromosome aberrations

Trisomic cells in short- and long-term CVS cultures have to be assessed carefully as they might not be confined to the placenta (confined placental mosaicism (CPM)); they can also represent the fetus. CPM is a common finding in CVS, occurring in 1–2% of cases18. Some chromosomes seem to be more prone to CPM than are others, and imprinted chromosomes have to be evaluated more carefully compared with others19. Therefore, the specific nature of the chromosome found to be trisomic must be considered in clinical decisions. Culture artifacts (such as single trisomic cells) must be distinguished from CPM. While the first almost never affects the fetus, and therefore has no clinical consequences, true CPM might cause IUGR. In contrast, a trisomic cell line in both short- and long-term cultures might suggest mosaicism in the fetus, and in such cases further evaluation by amniocentesis or cordocentesis will be performed. Mosaicism in either of these would indicate the fetus to be affected. For most chromosomes, there is no evidence that a more severe phenotype occurs in cases in which the trisomy mosaicism is combined with UPD in the normal cell line. Therefore, the phenotype in these cases seems to be caused basically by the trisomic cell line, and confirmation of an additional UPD is clinically irrelevant. If only a normal cell line is found on amniocentesis or cordocentesis, the question of UPD arises, but one has to keep in mind that mosaicism for some chromosomes (e.g. 17 and 20) might be found on CVS and/or amniocentesis, but not on cordocentesis, and the fetus might still be trisomic and clinically affected20. Testing for UPD is indicated if chromosome 14 or 15 is trisomic. For chromosomes 6, 7 and 16, the decision to perform prenatal testing is more difficult to make. Paternal UPD 6 or 7 has no direct clinical effect. Transient neonatal diabetes mellitus in paternal UPD 6 is treatable and diabetes mellitus type II, which might appear later in life, is a common disorder in western populations21 and should not be an indication for termination of pregnancy. Due to the small number of reported cases, knowledge of the long-term outcome of children with maternal UPD 7 is scarce, but psychomotor development in a good familial ambience as well as adult height after human growth hormone therapy usually lie within normal reference ranges. Most patients with maternal UPD 7 and a poor outcome have additional problems such as proven mosaicism or postnatal cerebral hemorrhage due to immaturity7. Maternal UPD 16 in the absence of malformations does not allow any prognosis regarding growth or psychomotor development. Furthermore, as discussed above, the likelihood of homozygosity for an autosomal recessively inherited mutation cannot be estimated. A trisomic cell line on CVS generally indicates formation by trisomy rescue subsequent to a meiotic error and therefore at least partial heterodisomy. Thus, as a general rule, prenatal UPD testing for chromosomes 6, 7 and 16 should be discussed critically on an individual basis. One exception might be if there is a known autosomal recessive disorder (eg. cystic fibrosis) in the family. However, in this case the parent(s) should first be tested for the mutation. Monosomy mosaicism rarely implies postnatal clinical problems, as cell lines monosomic for an autosomal chromosome are much less compatible with life than are trisomies.

Structural chromosome aberrations

In an otherwise unremarkable pregnancy, the question of UPD might also arise if an invasive prenatal diagnosis performed because of advanced maternal age or because of a known familial chromosomal aberration results in an abnormal karyotype being found. Because for chromosomes 13, 21 and 22, no genomic imprinting has been described, balanced homologous and non-homologous Robertsonian translocations between these chromosomes imply only the risk of hidden trisomy mosaicism or homozygosity for autosomal recessively inherited mutations. There is no evidence that the risk for hidden trisomy mosaicism is higher in the case of (heterodisomic) UPD than in the case of biparental inheritance. Therefore, UPD testing for these chromosomes is not indicated in routine prenatal diagnosis. Furthermore, a positive result might confuse the parents. In contrast, homologous and non-homologous Robertsonian translocations of the long arms of chromosomes 14 and/or 15, which carry imprinted genes, are an absolute indication to offer UPD testing. Although most homologous Robertsonian translocations are formed postzygotically22, the risk of UPD remains high. For both inherited and de novo non-homologous translocations, several recently published reviews demonstrate a low risk (c. 0.5%) for UPD23. Nevertheless, this figure justifies UPD testing in both situations, particularly if CVS or amniocentesis has been performed for other reasons. In the case of a parental non-homologous Robertsonian translocation, there might be a trisomy rescue in an initially trisomic zygote (i.e. 46,XN,t(13;14)+ 14) or a monosomy rescue after fertilization of a normal gamete by a nullisomic gamete, resulting in UPD from the parent not carrying the translocation. Several cases of trisomy rescue have been reported12 and one has to keep in mind that, due to hidden mosaicism, fetal anomalies are not completely excluded if biparental inheritance is found. If the translocation is de novo and formed meiotically, UPD should only be possible for the parent whose chromosomes are involved in the translocation.

Non-Robertsonian balanced translocations between any chromosomes are also only an indication for UPD testing if chromosomes carrying imprinted genes are involved. If chromosome 15 is involved and the breakpoint is around the Prader–Willi syndrome/Angelman syndrome critical region on 15q12, a deletion should first be excluded by FISH. One should keep in mind that such cases are very rare. Neither familial nor de novo balanced translocations between other chromosomes should be tested on a routine basis.

Another complex situation in prenatal diagnosis is the finding of a SMC. According to several studies, SMCs are found in 0.2–0.7% of all pregnancies and approximately half of them derive from chromosome 1519. Warburton24 and Crolla25 defined the risk of an abnormal phenotype in prenatally detected de novo SMCs as being 13% in general and 28% for non-acrocentric chromosomes. In practice, the chromosomal origin and familiarity or de novo occurrence must first be investigated before conclusions can be drawn. Demonstration of euchromatic material both in an additional inv dup(15) and in other SMCs indicates that congenital anomalies and mental retardation are likely and the question of UPD is secondary. However, if euchromatic material is excluded, the question of UPD is of great importance. In several cases, SMC(15) has been found in association with maternal and paternal UPD 1526, 27 and from several small studies, a risk of around 5% was estimated6. Therefore, testing for UPD 15 is indicated both in de novo and in familial cases. So far, only one case of SMC(14) with associated maternal UPD 14 has been described12. Nevertheless, testing is justified, although in the case of maternal UPD 14, genetic counseling is a great challenge due to the variable phenotype. For non-euchromatic SMC(6) or SMC(7), the situation is the same as for CPM of these chromosomes. For all other chromosomes, UPD testing is of academic value only and should not be performed in a routine setting.


The decision to perform prenatal UPD investigations depends mostly on the clinical situation, the chromosomal complement in question, and the phenotype associated with UPD of the particular chromosome. Although this sounds simple, in a specific situation the decision might not be easy. Furthermore, the decision to test for UPD should consider the possible result and particularly the clinical and psychological consequences for the parents. In principle, the decision approach can be chromosome-related or phenotype-related. In practice, both approaches will come together and should be accompanied by genetic counseling by an experienced clinical geneticist. To test all we can—sometimes primarily due to fear of prosecution—is certainly not the best way forward.


I thank Benno Röthlisberger, Center for Laboratory Medicine, Cantonal Hospital Aarau, Aarau, Switzerland for his very helpful comments.