American Journal of Medical Genetics Part C (Seminars in Medical Genetics) 169C:307–313 (2015)
R E V I E W A R T I C L E
What Every Clinical Geneticist Should Know About Testing for Osteogenesis Imperfecta in Suspected Child Abuse Cases MELANIE G. PEPIN
AND
PETER H. BYERS*
Non-accidental injury (NAI) is a major medical concern in the United States. One of the challenges in evaluation of children with unexplained fractures is that genetic forms of bone fragility are one of the differential diagnoses. Infants who present with fractures with mild forms of osteogenesis imperfecta (OI) (OI type I or OI type IV), the most common genetic form of bone disease leading to fractures might be missed if clinical evaluation alone is used to make the diagnosis. Diagnostic clinical features (blue sclera, dentinogenesis imperfecta, Wormian bones on X-rays or positive family history) may not be present or apparent at the age of evaluation. The evaluating clinician faces the decision about whether genetic testing is necessary in certain NAI cases. In this review, we outline clinical presentations of mild OI and review the history of genetic testing for OI in the NAI versus OI setting. We summarize our data of molecular testing in the Collagen Diagnostic Laboratory (CDL) from 2008 to 2014 where NAI was noted on the request for DNA sequencing of COL1A1 and COL1A2. We provide recommendations for molecular testing in the NAI versus OI setting. First, DNA sequencing of COL1A1, COL1A2, and IFITM5 simultaneously and duplication/deletion testing is recommended. If a causative variant is not identified, in the absence of a pathologic clinical phenotype, no additional gene testing is indicated. If a VUS is found, parental segregation studies are recommended. © 2015 Wiley Periodicals, Inc. KEY WORDS: osteogenesis imperfecta; child abuse; molecular testing osteogenesis imperfecta; biochemical testing osteogenesis imperfecta; recurrent fracture
How to cite this article: PEPIN MG, BYERS PH. 2015. What every clinical geneticist should know about testing for osteogenesis imperfecta in suspected child abuse cases. Am J Med Genet Part C Semin Med Genet 169C:307–313.
INTRODUCTION Non-accidental injury (NAI) is a major medical concern in the United States with more than 600,000 reports in 2013, [Children’s Bureau US Dept. Health & Human Services, 2013. Available at: http://www.acf.hhs.gov/programs/cb/ research-data-technology/statisticsresearch/child-maltreatment. Referenced July 20, 2015] of which a subset includes fractures. Roughly 15,000 children under the age of 36 months are hospitalized each year for fractures
[Leventhal et al., 2010]. Among these children the proportion of fractures
Abuse as a cause for fractures (without traumatic brain injury) is highest under the age of 12 months (20.4%), decreasing to 7.1% for 12 to 23 months of age and to 2.1% between 24 and 35 months. In
that population Leventhal and colleagues estimated that osteogenesis imperfecta (OI) accounted for 0.85% of the fractures. likely to have resulted from NAI in children hospitalized for fractures without traumatic brain injury is estimated to be 9.7%. Abuse as a cause for fractures (without traumatic brain injury) is
Melanie Pepin, M.S., C.G.C., is a board certified laboratory-based genetic counselor and researcher. Her career has included adult and pediatric genetic counseling, genetic testing, teaching, and mentoring of students. Melanie has more than 40 research publications that focus on the translation of laboratory research to the genetic counseling clinic setting for inherited connective tissue disorders. Dr. Peter Byers is Director of the Collagen Diagnostic Laboratory (CDL) and Director of the Center for Precision Diagnostics at the University of Washington (UW) in Seattle, WA. He is Professor of Medicine and Pathology, and Adjunct Professor of Genome Sciences at UW. He is actively involved in several lines of research including the characterization of mutations in genes that give rise to forms of osteogenesis imperfecta and other disorders, identification of new OI-related genes, and the identification and characterization of mechanisms of action of mutations in genes that give rise to other heritable disorders, including those that lead to aneurysm and dissection. *Correspondence to: Peter H. Byers, M.D., University of Washington-Pathology, Box 357470 Seattle Washington 98195-7470, United States. E-mail: [email protected] DOI 10.1002/ajmg.c.31459 Article first published online 14 November 2015 in Wiley Online Library (wileyonlinelibrary.com).
ß 2015 Wiley Periodicals, Inc.
REVIEW ARTICLE
highest under the age of 12 months (20.4%), decreasing to 7.1% for 12 to 23 months of age and to 2.1% between 24 and 35 months [Leventhal et al., 2010]. In that population Leventhal and colleagues estimated that osteogenesis imperfecta (OI) accounted for 0.85% of the fractures. One of the concerns in evaluation of children with unexplained fractures is that genetic forms of bone fragility are one of the differential diagnoses. The most common genetic form of bone disease that results in fractures in childhood is OI [Bronicki et al., 2015]. The incidence of OI has been estimated to be 1/10,000 [Sillence et al., 1979; Stoll et al., 1989] of which roughly 85% were thought to have nonlethal forms. Since those surveys, the incidence of lethal forms of OI has diminished as a result of early pregnancy screening and directed prenatal diagnosis so that non-lethal forms currently appear to constitute more than 95% of children now born with OI and about 400 births per year in the US. If each of the children with non-lethal OI in the 0–3 year cohort were hospitalized because of “unexplained fractures“ (which is clearly an overestimate) then they would comprise about 1,200 of the 15,000 children (12%) hospitalized for this reason. Leventhal et al., suggested that OI explained fractures in about 1% of the group [Leventhal et al., 2010]. However, the possibility remains that infants with mild forms of OI (OI type I or OI type IV) might be missed if clinical evaluation alone is used to make the diagnosis. They may not be identified because some of the diagnostic clinical features (blue sclera, dentinogenesis imperfecta or Wormian bones on Xrays) may not be present or apparent at those ages and because some children with OI do not have a family history of the disorder [Marlowe et al., 2002]. As a result the evaluating clinician and often the clinical geneticist face the decision about when genetic testing is necessary to come to a diagnosis of OI. Regardless of the decision regarding genetic testing, it is important to note that there are mandatory child abuse reporting laws for health care providers in the US, and
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
medical professionals should not hesitate to report concern for child abuse even if components of the medical work-up are pending or ongoing [Flaherty et al., 2014]. In this review, we outline clinical presentations of mild OI and review the history of genetic testing for OI in the NAI versus OI setting. We summarize our recent data regarding the percentage of OI diagnoses in a review of molecular test requests submitted to the Collagen Diagnostic Laboratory (CDL) from 2008 to 2014 where NAI was noted on the request for DNA sequence analysis of COL1A1 and COL1A2 for 305 samples. Finally, we conclude with our latest recommendations for molecular testing in the NAI versus OI setting.
CLINICAL IDENTIFICATION OF CHILDREN WITH GENETIC BASES FOR FRACTURES Can we identify the infants and toddlers with OI in whom the diagnosis goes unnoticed until a fracture occurs that raises the question of non-accidental injury? The child with the following mild forms of OI and a negative family history may be missed clinically especially before 36 months of age.
OI TYPE I OI type I results from heterozygous null mutations in COL1A1. Although some children with OI type I are identified by prenatal ultrasound findings (bowed femurs or, less often, fracture) or blue sclera at birth, the majority are born fracture-free with a normal exam. In the absence of a family history, bluish sclera may be considered a normal age-related variant commonly found until 6–9 months. Teeth are generally normal and long bone fractures may not occur until children begin to walk.
OI TYPE IV OI type IV results from heterozygous missense mutations in either COL1A1 or COL1A2. De novo mutations are
308
frequent. [Patel et al., 2015] Many infants with OI type IV are fracture free at birth and have normal sclerae. Children with OI type IV may be small for their age and develop long bone bowing and fractures in childhood, but during the first two years of life, even in a family with OI, genetic testing might be required to confirm the diagnosis.
OI TYPE V OI type V results from a single mutation in IFITM5 that results in the addition of five amino acids at the amino terminal end of the encoded protein, BRIL (Bone-restricted interferon-induced transmembrane protein-like protein, which, however, is not interferon induced). The natural history of OI type V is characterized by fractures, short stature, normal sclerae, reduced bone density, hypertrophic callus and bony hypertrophy, of the radius and ulna. In a child without a family history, the diagnosis might be missed because of the age-depended, and variable expression of the disorder [Shapiro et al., 2013]. There is often a mesh-like pattern of lamellation in bone when examined under polarized light microscopy [Rauch et al., 2013]. The periosteal hyperostosis and new bone formation is often described as mineralization of the interosseous membrane between the ulna and radius and can be seen by 6 months of age [Arundel et al., 2011]. The exact mechanism of disease and the explanation for variability of the phenotype are under investigation. Until the full IFITM5 OI type V phenotypic spectrum is clear, sequencing of the gene is necessary in testing [Grover et al., 2013; Guillen-Navarro et al., 2014].
RARE RECESSIVE FORMS OF OI There are rare recessive OI phenotypes that might be overlooked in the absence of a thorough genetic evaluation OI type VI, which results from biallelic mutations in SERPINF1, often presents with few if any signs of OI at birth and the infants begin to develop progressive
309
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
scoliosis and bone deformity by about a year of age. There are few clinical clues to this diagnosis until bone deformity develops and so parental consanguinity may be the trigger for specific diagnostic studies. Bruck syndrome is a genetically
OI type VI, which results from biallelic mutations in SERPINF1, often presents with few if any signs of OI at birth, and the infants begin to develop progressive scoliosis and bone deformity by about a year of age. There are few clinical clues to this diagnosis until bone deformity develops and so parental consanguinity may be the trigger for specific diagnostic studies. heterogeneous recessively inherited disorder that results from biallelic mutations in PLOD2 or in FKBP10 [Alanay et al., 2010; Ha-Vinh et al., 2004; Schwarze et al., 2013]. Most infants with these conditions have congenital contractures, but in their absence fractures can raise concern about NAI. In this context genetic diagnosis is the usual pathway to diagnosis.
CLINICAL SIGNS THAT MAY HELP TO DISTINGUISH CHILDREN WHO HAVE BEEN ABUSED FROM THOSE WITH MILD OI The work-up for suspected physical abuse includes physical examination, radiological assessmen, and laboratory studies [Flaherty et al., 2014]. In some cases, children may have findings that can be seen with both child abuse and underlying genetic conditions. In other cases, children may have injuries that are highly suspicious for abuse, such as
bruising in the pattern of an object or hand [Anderst et al., 2013]. Posterior rib fractures, transverse fractures of several digits, metacarpals or metatarsals, and metaphyseal chip fractures were considered strongly suggestive of NAI when Kempe made his initial observations [Kempe et al., 1984]. Although each of these types of fracture can be seen in infants and children with OI, their frequency remains substantially greater in the context of trauma [Greeley et al., 2013; Shelmerdine et al., 2015]. Given the clinical importance of the distinction
Posterior rib fractures, transverse fractures of several digits, metacarpals or metatarsals, and metaphyseal chip fractures were considered strongly suggestive of NAI when Kempe made his initial observations . Although each of these types of fracture can be seen in infants and children with OI, their frequency remains substantially greater in the context of trauma. between abuse and biological causes of fracture, the recognition that no fracture type is pathognomonic of NAI has led to the search for biological tests that could distinguish between the two options.
HISTORY OF GENETIC TESTING FOR OI In the early 1970s, it was found that cultured dermal fibroblasts from individuals with different forms of OI usually made less type I procollagen than cells from controls or the chains of type I procollagen had altered electrophoretic mobilities. These tests were developed to facilitate the diagnosis of OI and to try to differentiate among individuals with different clinical presentations. Very quickly this type of
REVIEW ARTICLE
study began to be used to try to exclude the diagnosis of OI in the context of fractures thought to be related to NAI. This led Wenstrup and colleagues to ask if this test could identify abnormalities in cells from all individuals with well documented OI [Wenstrup et al., 1990]. They found that among those with non-lethal forms of OI, cells from about 87% of these individuals either made too little collagen or the collagen they made behaved abnormally when separated on polyacrylamide gels. This finding led defense attorneys to argue in criminal cases that a negative result did not exclude the diagnosis of OI “beyond a reasonable doubt.” This argument assumed that the a priori probability that a child who was tested had OI was very high, something that was rarely correct. In 1996, Steiner et al. [1996] reviewed studies of collagen production by cultured fibroblasts from 48 children who were tested to distinguish OI from NAI. Of these, six had abnormal studies consistent with OI (12.5%), but only one patient (2% of the total) had not been suspected to have OI, based on clinical findings, by the time the tests were completed. Marlowe et al., in 2002 examined the results of the same type of studies in an additional 262 children with a differential diagnosis that included NAI and OI [Marlowe et al., 2002]. They found 11 (4.1%) with a type I collagen protein abnormality consistent with OI. This study confirmed that clinical examination alone did not identify some children with mild forms of OI [Marlowe et al., 2002]. Review of clinical data from 138 children in that study, for whom sufficient clinical information accompanied the samples submitted to the CDL, revealed that in six instances the diagnosis of OI was suspected on clinical grounds by the time studies were completed. Still, reviewing all data from all the children identified in that study to have OI by this type of testing, no combination of additional clinical signs (for example, blue sclerae, short stature), associated history (family history or history of injury), or radiological features (nature and location of the fractures or Wormian
REVIEW ARTICLE
bones or bowed bones), identified all affected infants. Overall, the incidence of OI identified by analysis of collagen production by cultured fibroblasts among children evaluated for NAI was 5%, and those missed based on clinical or radiological features was about 2%.
Overall, the incidence of OI identified by analysis of collagen production by cultured fibroblasts among children evaluated for NAI was 5%, and those missed based on clinical or radiological features was about 2%. Pathogenic alterations in COL1A1 or COL1A2 account for 90–95% of individuals with non-lethal forms of OI sequenced in the CDL and about the same reported in the Database of OI Mutations (http://www.le.ac.uk/ genetics/collagen/; CDL Data). A subset of pathogenic variants that lead to changes in the amino terminal regions of the proa chains of type I procollagen escape recognition by protein screening alone [Cabral et al., 2005; Marini et al., 2007]. Because DNA-based testing more accurately identifies individuals with mild OI, analysis of collagens produced by cultured fibroblasts has taken a back seat to molecular genetic testing. Presently, genomic DNA sequence analysis of both COL1A1 and COL1A2 is the first step in laboratory testing when the diagnosis of OI is considered in most patients.
HOW DOES DNA SEQUENCE ANALYSIS COMPARE WITH PROTEIN STUDIES FOR DETECTION OF OI IN THE OI V NAI SETTING? In a recent review of molecular test requests submitted to the CDL from 2008 to 2014, there were 305 samples in
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
which NAI was noted on the request for DNA sequencing of COL1A1 and COL1A2. Ninety percent (n ¼ 276) of the children were under the age of 1 year (1–60 months) with a mean age of testing of 7 months (median 5 months). Nine children (3% of studied samples) had a pathogenic variant in COL1A1 (6) or COL1A2 (3), a frequency similar to the reported frequency in protein based testing. Two of the nine infants were also suspected to have OI by clinical evaluation at the time of testing. Thus, 7/9 infants found to have OI (2% of the total sample) were missed by clinical evaluation alone. Review of limited clinical details confirmed that those identified with OI could not be clinically distinguished (on paper) from those with normal test results. One important issue in this evaluation is that the cohort tested is a highly selected subset of all those infants and children evaluated for NAI. In most instances, the criteria used to select children to be tested remain elusive. A concern with molecular testing is that a variant of unknown significance (VUS) could be found; in most settings, when the a priori likelihood to find a pathogenic variant is low, the chance to find a VUS is the same as in the general population. A VUS was identified in 20 (6.5%) of 305 children undergoing COL1A1 and COL1A2 sequencing in this setting. Of these 12 were identified in COL1A1 and eight in COL1A2. When a VUS is identified, the usual request is to study parents (or other family members) to determine if the variant segregates with the phenotype in question. This request may lead to determination that the variant was inherited from an individual with no bone phenotype and strengthen the case against the parents—a form of selfincrimination. In this setting the ethical, legal, and social issues may prove challenging.
RECOMMENDATION FOR LABORATORY TESTING IN OI V NAI SETTING Laboratory guidelines for the diagnosis of OI were developed by the European
310
Molecular Genetics Quality Network (EMQN) on the basis of the experience of 17 OI expert clinicians and diagnostic laboratory directors [van et al., 2012]. In the setting of suspected NAI in the absence of clear phenotypic signs of OI or a family history, using EMQN guidance, testing would truncate after type I collagen gene sequence analysis was completed (See Fig. 1 from van et al., 2012). The suggested guidelines in the OI v NAI setting include analysis of IFITM5 sequence concurrently.
SUGGESTED APPROACH TO TESTING FOR OI IN THE CONTEXT OF CONCERNS ABOUT NAI At this time there are no consensus guidelines that identify the circumstances in which genetic testing for OI is indicated or when clinical and radiological evaluation alone can exclude the diagnosis of OI. Additional research on number of fractures, timing of fractures, and types of fractures seen among patients presenting with OI could help to establish definitive guidelines for molecular testing. Testing for OI can be completed as follows: 1. Sequence COL1A1, COL1A2, and
IFITM5 simultaneously. Exclude deletion or duplication of part or all of these three genes by high resolution array CGH or MLPH duplication/deletion testing. If a causative variant is not identified, in the absence of a significant pathologic clinical phenotype, no additional gene testing is indicated. Sixteen (16) genes are presently known to play a role in fracture susceptibility in children with recessively inherited forms of OI [Byers and Pyott, 2012; Cho et al., 2012; Lapunzina et al., 2010; Martınez-Glez et al., 2012; Semler et al., 2012; Shaheen et al., 2012; Pyott et al., 2013; Symoens et al., 2013; Garbes et al., 2015; Mendoza-Londono et al., 2015; Rauch et al., 2015]. In total, these account for less than 10% of all individuals identified with OI. With rare exceptions, the infants with recessively inherited forms of OI
311
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
Figure 1.
REVIEW ARTICLE
Recommended approach to genetic testing for osteogenesis imperfecta in a child with signs of non-accidental injury.
usually present with prenatal or neonatal onset of fractures, and moderate to severe bony phenotype. Second tier testing for recessively inherited forms of OI is not recommended in instances of unexplained fractures with or without blue sclera unless other features, such as consanguinity or congenital contractures, demand reconsideration. 2. If a variant of uncertain significance (VUS) is identified, complete recommended segregation studies (see discussion above). The likelihood to find a VUS in the absence of a known pathogenic variant in the type I collagens genes is 6.5% (CDL data). The VUS may be a synonymous change, an alteration in intron sequence or result in a non-glycine substitution in the triple helical
domain. The non-glycine substitutions may occur in the X or Y-positions of the triple helical domain that is characterized by the repeating Gly-Xaa-Yaa triplet that extends for 1,014 residues. Segregation studies are the first step in following up the uncertain test outcome. Of the 20 VUS reported to an ordering physician in the 2008–2014 CDL review, there were eight instances of follow-up testing of one or both parents (40%) an indication that testing recommended for final interpretation is often ignored, misunderstood, or not undertaken because of concern for the implications of such studies, as noted above. There are several obstacles to parental testing (perceived costs, parental availability, and removal of child from the family) that may also limit further tests.
Clinical evaluation for child abuse includes consideration of differential diagnoses such as accidental trauma, and underlying bone disease. In addition to a careful medical history and physical examination, the American Academy of Pediatrics report discusses laboratory, and radiology testing to assess for occult injury and bone disease [Flaherty et al., 2014]. Testing for occult injury can include liver function studies, skeletal survey and head imaging with decisions based on clinical presentation and patient age. Genetic tests, like those for OI, are targeted specifically and the diagnostic and exclusion value depend on test sensitivity and specificity, and the a priori likelihood that a child in the test environment could have OI. For example, if the a priori likelihood to have OI is 2% (based on chance of mild OI
REVIEW ARTICLE
without any clinical features or family history in NAI cases) and the test identifies 95% of people with OI, then an infant with a normal test has an approximately 1/980 chance to have OI as a consequence of a pathogenic variant in a type I collagen gene. Historically,
If the a priori likelihood to have OI is 2% (based on chance of mild OI without any clinical features or family history in NAI cases) and the test identifies 95% of people with OI, then an infant with a normal test has an approximately 1/980 chance to have OI as a consequence of a pathogenic variant in a type I collagen gene. this type of calculation has been avoided in laboratory results because of less certainty about test sensitivity, and the recognition that OI is not the only alternative explanation to NAI.
CONCLUSION As we pointed out in the Introduction, if all children with non-lethal OI under the age of 3 years had fractures that led to evaluation of NAI, they would represent a very small subset of the children evaluated. There are many tools to facilitate the diagnosis of OI and these include evaluation of family history, careful consideration of the context of fracture, an informed clinical examination, and radiographic study of the fracture and other bones. Even with these tools, evaluation of a small subset (about 5% as per our laboratory data) of the children in whom a diagnosis of NAI is considered identified biochemical or genetic alterations consistent with OI. Given the high social cost of a missed diagnosis in these children, the current practice of testing appears to be justified.
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
One of the major issues in the current selection of children to be tested is the criteria used. The number of children tested represents a very small subset of those considered each year to be at risk for NAI. It is clear that for some children the clinical tools are sufficient to make the diagnosis and testing is deferred for a later time when genetic counseling for the family, or evaluation of reproductive choices for the child might be considered. Careful evaluation of this decision process could evolve more powerful clinical mechanisms that would identify more children with OI. Currently, recommendations include DNA determination of the sequence of COL1A1, COL1A2, and IFITM5 simultaneously and exclude deletion or duplication of part or all of these three genes (array CGH or MLPH duplication/deletion testing). If a causative variant is not identified, in the absence of a significant pathologic clinical phenotype, no additional gene testing is indicated. If a VUS is found, parental segregation studies are recommended but may be hard to obtain due to follow-up or other social issues. Future studies on implications of VUS in NAI cases may offer guidance to clinicians when this situation arises. Meanwhile, in these complex cases, the molecular laboratory serves to provide a piece of puzzle. Interpreting results and implications for each child, and considering the diagnosis of NAI versus OI, ultimately involves a number of experts; evaluations are multidisciplinary and may involve medical providers, child protective services, and in some cases law enforcement professionals.
REFERENCES Alanay Y, Avaygan H, Camacho N, Utine G, Boduroglu K, Aktas D, Alikasifoglu M, Tuncbilek E, Orhan D, Bakar F, Zabel B, Superti-Furga A, Bruckner-Tuderman L, Curry C, Pyott S, Byers P, Eyre D, Baldridge D, Lee B, Merrill A, Davis E, Cohn D, Akarsu N, Krakow D. 2010. Mutations in the gene encoding the RER protein FKBP65 cause autosomal-recessive osteogenesis imperfecta. Am J Hum Genet 86:551–559. Anderst JD, Carpenter SL, Abshire TC. 2013. Section on Hematology/Oncology and
312
committee on child abuse and neglect of the American Academy of Pediatrics. Pediatrics 131:e1314–e1322. Arundel P, Offiah A, Bishop NJ. 2011. Evolution of the radiographic appearance of the metaphyses over the first year of life in type V osteogenesis imperfecta: Clues to pathogenesis. J Bone Miner Res 26: 894–898. Bronicki L, Stevenson RE, Spranger J. 2015. Beyond osteogenesis imperfecta: Causes of fractures in childhood. AJMG. Byers PH, Pyott SM. 2012. Recessively inherited forms of osteogenesis imperfecta. Annu Rev Genet 46:475–497. Cabral WA, Makareeva E, Colige A, Letocha AD, Ty JM, Yeowell HN, Pals G, Leikin S, Marini JC. 2005. Mutations near amino end of alpha1(I) collagen cause combined osteogenesis imperfecta/Ehlers-Danlos syndrome by interference with N-propeptide processing. J Biol Chem 280:19259–19269. Cho TJ, Lee KE, Lee SK, Song SJ, Kim KJ, Jeon D, Lee G, Kim HN, Lee HR, Eom HH, Lee ZH, Kim OH, Park WY, Park SS, Ikegawa S, Yoo WJ, Choi IH, Kim JW. 2012. A single recurrent mutation in the 50 -UTR of IFITM5 causes osteogenesis imperfecta type V. Am J Hum Genet 91:343–348. Flaherty EG, Perez-Rossello JM, Levine MA, Hennrikus WL, America acamady of pediatric committee on child abuse and neglect, section radiology, section on endocrinology, section on orthopaedics and the society for pediatric radiology. 2014. Evaluating children with fractures for child physical abuse. Pediatrics 133:e477–e489. Garbes L, Kim K, Rieß A, Hoyer-Kuhn H, Beleggia F, Bevot A, Kim MJ, Huh YH, Kweon HS, Savarirayan R, Amor D, Kakadia PM, Lindig T, Kagan KO, Becker J, Boyadjiev SA, Wollnik B, Semler O, Bohlander SK, Kim J, Netzer C. 2015. Mutations in SEC24D, encoding a component of the COPII machinery, cause a syndromic form of osteogenesis imperfecta. Am J Hum Genet 96:432–439. Greeley CS, Donaruma-Kwoh M, Vettimattam M, Lobo C, Williard C, Mazur L. 2013. Fractures at diagnosis in infants and children with osteogenesis imperfecta. J Pediatr Orthop 33:32–36. Grover M, Campeau PM, Lietman CD, Lu JT, Gibbs RA, Schlesinger AE, Lee BH. 2013. Osteogenesis imperfecta without features of type V caused by a mutation in the IFITM5 gene. J Bone Miner Res 28:2333–2337. Guillen-Navarro E, Ballesta-Martınez MJ, Valencia M, Bueno AM, Martinez-Glez V, L opezGonzalez V, Burnyte B, Utkus A, Lapunzina P, Ruiz-Perez VL. 2014. Two mutations in IFITM5 causing distinct forms of osteogenesis imperfecta. Am J Med Genet A 164A:1136–1142. Ha-Vinh R, Bank AY, Campos-Xavier RA, Zankl AB, Superti-Furga A, Bonafe A. 2004. Phenotypic and molecular characterization of Bruck syndrome (osteogenesis imperfecta with contractures of the large joints) caused by a recessive mutation in PLOD2. Am J Med Genet A 131:115–120. Kempe CH, Silverman FN, Steele BF, Droegemueller W, Silver HK. 1962. Landmark article July 7, 1962: The battered-child
313
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
syndrome. By C. Henry Kempe, Frederic N. Silverman, Brandt F. Steele, William Droegemueller, and Henry K. Silver. JAMA 251:3288–3294. Lapunzina P, Aglan M, Temtamy S, Caparr osMartın J, Valencia M, Let on R, MartınezGlez V, Elhossini R, Amr K, Vilaboa N, Ruiz-Perez V. 2010. Identification of a frameshift mutation in osterix in a patient with recessive osteogenesis imperfecta. Am J Hum Genet 87:110–114. Leventhal JM, Martin KD, Asnes AG. 2010. Fractures and traumatic brain injuries: Abuse versus accidents in a US database of hospitalized children. Pediatrics 126: e104–e115. Marini JC, Forlino A, Cabral WA, Barnes AM, San Antonio JD, Milgrom S, Hyland JC, K€ orkk€ o J, Prockop DJ, De Paepe A, Coucke P, Symoens S, Glorieux FH, Roughley PJ, Lund AM, Kuurila-Svahn K, Hartikka H, Cohn DH, Krakow D, Mottes M, Schwarze U, Chen D, Yang K, Kuslich C, Troendle J, Dalgleish R, Byers PH. 2007. Consortium for osteogenesis imperfecta mutations in the helical domain of type I collagen: Regions rich in lethal mutations align with collagen binding sites for integrins and proteoglycans. Hum Mutat 28:209–221.Review. Marlowe A, Pepin M, Byers P. 2002. Testing for osteogenesis imperfecta in cases of suspected non-accidental injury. J Med Genet 39:382–386. Martınez-Glez V, Valencia M, Caparr os-Martın JA, Aglan M, Temtamy S, Tenorio J, Pulido V, Lindert U, Rohrbach M, Eyre D, Giunta C, Lapunzina P, Ruiz-Perez VL. 2012. Identification of a mutation causing deficient BMP1/mTLD proteolytic activity in autosomal recessive osteogenesis imperfecta. Hum Mutat 33:343–350. Mendoza-Londono R, Fahiminiya S, Majewski J, Tetreault M, Nadaf J, Kannu P, Sochett E, Howard A, Stimec J, Dupuis L, Roschger P, Klaushofer K, Palomo T, Ouellet J, Al-Jallad H, Mort JS, Moffatt P, Boudko S, B€achinger HP, Rauch F, Consortium CRC. 2015. Recessive osteogenesis imperfecta caused by missense mutations in SPARC. Am J Hum Genet 96:979–985.
Patel RM, Nagamani SC, Cuthbertson D, Campeau PM, Krischer JP, Shapiro JR, Steiner RD, Smith PA, Bober MB, Byers PH, Pepin M, Durigova M, Glorieux FH, Rauch F, Lee BH, Hart T, Sutton VR. 2015. A cross-sectional multicenter study of osteogenesis imperfecta in North Americaresults from the linked clinical research centers. Clin Genet 87:133–140. Pyott SM, Tran TT, Leistritz DF, Pepin MG, Mendelsohn NJ, Temme RT, Fernandez BA, Elsayed SM, Elsobky E, Verma I, Nair S, Turner EH, Smith JD, Jarvik GP, Byers PH. 2013. WNT1 Mutations in families affected by moderately severe and progressive recessive osteogenesis imperfecta. Am J Hum Genet 92:590–597. Rauch F, Fahiminiya S, Majewski J, Carrot-Zhang J, Boudko S, Glorieux F, Mort JS, B€achinger HP, Moffatt P. 2015. Cole-Carpenter syndrome is caused by a heterozygous missense mutation in P4HB. Am J Hum Genet 96:425–431. Rauch F, Moffatt P, Cheung M, Roughley P, Lalic L, Lund AM, Ramirez N, Fahiminiya S, Majewski J, Glorieux FH. 2013. Osteogenesis imperfecta type V: Marked phenotypic variability despite the presence of the IFITM5 c.-14C>T mutation in all patients. J Med Genet 50:21–24. Schwarze U, Cundy T, Pyott SM, Christiansen HE, Hegde MR, Bank RA, Pals G, Ankala A, Conneely K, Seaver L, Yandow SM, Raney E, Babovic-Vuksanovic D, Stoler J, Ben-Neriah Z, Segel R, Lieberman S, Siderius L, Al-Aqeel A, Hannibal M, Hudgins L, McPherson E, Clemens M, Sussman MD, Steiner RD, Mahan J, Smith R, Anyane-Yeboa K, Wynn J, Chong K, Uster T, Aftimos S, Sutton VR, Davis EC, Kim LS, Weis MA, Eyre D, Byers PH. 2013. Mutations in FKBP10, which result in Bruck syndrome and recessive forms of osteogenesis imperfecta, inhibit the hydroxylation of telopeptide lysines in bone collagen. Hum Mol Genet 22:1–17. Semler O, Garbes L, Keupp K, Swan D, Zimmermann K, Becker J, Iden S, Wirth B, Eysel P, Koerber F, Schoenau E, Bohlander SK, Wollnik B, Netzer C. 2012.
REVIEW ARTICLE A mutation in the 5’-UTR of IFITM5 creates an in-frame start codon and causes autosomal-dominant osteogenesis imperfecta type V with hyperplastic callus. Am J Hum Genet 91:349–357. Shaheen R, Alazami AM, Alshammari MJ, Faqeih E, Alhashmi N, Mousa N, Alsinani A, Ansari S, Alzahrani F, Al-Owain M, Alzayed ZS, Alkuraya FS. 2012. Study of autosomal recessive osteogenesis imperfecta in Arabia reveals a novel locus defined by TMEM38B mutation. J Med Genet 49:630–635. Shapiro JR, Lietman C, Grover M, Lu JT, Nagamani SC, Dawson BC, Baldridge DM, Bainbridge MN, Cohn DH, Blazo M, Roberts TT, Brennen FS, Wu Y, Gibbs RA, Melvin P, Campeau PM, Lee BH. 2013. Phenotypic variability of osteogenesis imperfecta type V caused by an IFITM5 mutation. J Bone Miner Res 28:1523– 1530. Shelmerdine SC, Arthurs OJ, Calder AD. 2015. Rib fractures in osteogenesis imperfecta: Have we learnt anything about child abuse? J Pediatr Orthop 35:e81. Sillence D, Senn A, Danks D. 1979. Genetic heterogeneity in osteogenesis imperfecta. J Med Genet 16:101–116. Stoll C, Dott B, Roth M, Alembik Y. 1989. Birth prevalence rates of skeletal dysplasias. Clin Genet 35:88–92. Symoens S, Malfait F, D’hondt S, Callewaert B, Dheedene A, Steyaert W, B€achinger HP, De Paepe A, Kayserili H, Coucke PJ. 2013. Deficiency for the ER-stress transducer OASIS causes severe recessive osteogenesis imperfecta in humans. Orphanet J Rare Dis 8:154. van Dijk FS, Byers PH, Dalgleish R, Malfait F, Maugeri A, Rohrbach M, Symoens S, Sistermans EA, Pals G. 2012. EMQN best practice guidelines for the laboratory diagnosis of osteogenesis imperfecta. Eur J Hum Genet 20:11–19. Wenstrup R, Willing M, Starman B, Byers P. 1990. Distinct biochemical phenotypes predict clinical severity in nonlethal variants of osteogenesis imperfecta. Am J Hum Genet 46:975–982.
R E V I E W A R T I C L E
What Every Clinical Geneticist Should Know About Testing for Osteogenesis Imperfecta in Suspected Child Abuse Cases MELANIE G. PEPIN
AND
PETER H. BYERS*
Non-accidental injury (NAI) is a major medical concern in the United States. One of the challenges in evaluation of children with unexplained fractures is that genetic forms of bone fragility are one of the differential diagnoses. Infants who present with fractures with mild forms of osteogenesis imperfecta (OI) (OI type I or OI type IV), the most common genetic form of bone disease leading to fractures might be missed if clinical evaluation alone is used to make the diagnosis. Diagnostic clinical features (blue sclera, dentinogenesis imperfecta, Wormian bones on X-rays or positive family history) may not be present or apparent at the age of evaluation. The evaluating clinician faces the decision about whether genetic testing is necessary in certain NAI cases. In this review, we outline clinical presentations of mild OI and review the history of genetic testing for OI in the NAI versus OI setting. We summarize our data of molecular testing in the Collagen Diagnostic Laboratory (CDL) from 2008 to 2014 where NAI was noted on the request for DNA sequencing of COL1A1 and COL1A2. We provide recommendations for molecular testing in the NAI versus OI setting. First, DNA sequencing of COL1A1, COL1A2, and IFITM5 simultaneously and duplication/deletion testing is recommended. If a causative variant is not identified, in the absence of a pathologic clinical phenotype, no additional gene testing is indicated. If a VUS is found, parental segregation studies are recommended. © 2015 Wiley Periodicals, Inc. KEY WORDS: osteogenesis imperfecta; child abuse; molecular testing osteogenesis imperfecta; biochemical testing osteogenesis imperfecta; recurrent fracture
How to cite this article: PEPIN MG, BYERS PH. 2015. What every clinical geneticist should know about testing for osteogenesis imperfecta in suspected child abuse cases. Am J Med Genet Part C Semin Med Genet 169C:307–313.
INTRODUCTION Non-accidental injury (NAI) is a major medical concern in the United States with more than 600,000 reports in 2013, [Children’s Bureau US Dept. Health & Human Services, 2013. Available at: http://www.acf.hhs.gov/programs/cb/ research-data-technology/statisticsresearch/child-maltreatment. Referenced July 20, 2015] of which a subset includes fractures. Roughly 15,000 children under the age of 36 months are hospitalized each year for fractures
[Leventhal et al., 2010]. Among these children the proportion of fractures
Abuse as a cause for fractures (without traumatic brain injury) is highest under the age of 12 months (20.4%), decreasing to 7.1% for 12 to 23 months of age and to 2.1% between 24 and 35 months. In
that population Leventhal and colleagues estimated that osteogenesis imperfecta (OI) accounted for 0.85% of the fractures. likely to have resulted from NAI in children hospitalized for fractures without traumatic brain injury is estimated to be 9.7%. Abuse as a cause for fractures (without traumatic brain injury) is
Melanie Pepin, M.S., C.G.C., is a board certified laboratory-based genetic counselor and researcher. Her career has included adult and pediatric genetic counseling, genetic testing, teaching, and mentoring of students. Melanie has more than 40 research publications that focus on the translation of laboratory research to the genetic counseling clinic setting for inherited connective tissue disorders. Dr. Peter Byers is Director of the Collagen Diagnostic Laboratory (CDL) and Director of the Center for Precision Diagnostics at the University of Washington (UW) in Seattle, WA. He is Professor of Medicine and Pathology, and Adjunct Professor of Genome Sciences at UW. He is actively involved in several lines of research including the characterization of mutations in genes that give rise to forms of osteogenesis imperfecta and other disorders, identification of new OI-related genes, and the identification and characterization of mechanisms of action of mutations in genes that give rise to other heritable disorders, including those that lead to aneurysm and dissection. *Correspondence to: Peter H. Byers, M.D., University of Washington-Pathology, Box 357470 Seattle Washington 98195-7470, United States. E-mail: [email protected] DOI 10.1002/ajmg.c.31459 Article first published online 14 November 2015 in Wiley Online Library (wileyonlinelibrary.com).
ß 2015 Wiley Periodicals, Inc.
REVIEW ARTICLE
highest under the age of 12 months (20.4%), decreasing to 7.1% for 12 to 23 months of age and to 2.1% between 24 and 35 months [Leventhal et al., 2010]. In that population Leventhal and colleagues estimated that osteogenesis imperfecta (OI) accounted for 0.85% of the fractures. One of the concerns in evaluation of children with unexplained fractures is that genetic forms of bone fragility are one of the differential diagnoses. The most common genetic form of bone disease that results in fractures in childhood is OI [Bronicki et al., 2015]. The incidence of OI has been estimated to be 1/10,000 [Sillence et al., 1979; Stoll et al., 1989] of which roughly 85% were thought to have nonlethal forms. Since those surveys, the incidence of lethal forms of OI has diminished as a result of early pregnancy screening and directed prenatal diagnosis so that non-lethal forms currently appear to constitute more than 95% of children now born with OI and about 400 births per year in the US. If each of the children with non-lethal OI in the 0–3 year cohort were hospitalized because of “unexplained fractures“ (which is clearly an overestimate) then they would comprise about 1,200 of the 15,000 children (12%) hospitalized for this reason. Leventhal et al., suggested that OI explained fractures in about 1% of the group [Leventhal et al., 2010]. However, the possibility remains that infants with mild forms of OI (OI type I or OI type IV) might be missed if clinical evaluation alone is used to make the diagnosis. They may not be identified because some of the diagnostic clinical features (blue sclera, dentinogenesis imperfecta or Wormian bones on Xrays) may not be present or apparent at those ages and because some children with OI do not have a family history of the disorder [Marlowe et al., 2002]. As a result the evaluating clinician and often the clinical geneticist face the decision about when genetic testing is necessary to come to a diagnosis of OI. Regardless of the decision regarding genetic testing, it is important to note that there are mandatory child abuse reporting laws for health care providers in the US, and
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
medical professionals should not hesitate to report concern for child abuse even if components of the medical work-up are pending or ongoing [Flaherty et al., 2014]. In this review, we outline clinical presentations of mild OI and review the history of genetic testing for OI in the NAI versus OI setting. We summarize our recent data regarding the percentage of OI diagnoses in a review of molecular test requests submitted to the Collagen Diagnostic Laboratory (CDL) from 2008 to 2014 where NAI was noted on the request for DNA sequence analysis of COL1A1 and COL1A2 for 305 samples. Finally, we conclude with our latest recommendations for molecular testing in the NAI versus OI setting.
CLINICAL IDENTIFICATION OF CHILDREN WITH GENETIC BASES FOR FRACTURES Can we identify the infants and toddlers with OI in whom the diagnosis goes unnoticed until a fracture occurs that raises the question of non-accidental injury? The child with the following mild forms of OI and a negative family history may be missed clinically especially before 36 months of age.
OI TYPE I OI type I results from heterozygous null mutations in COL1A1. Although some children with OI type I are identified by prenatal ultrasound findings (bowed femurs or, less often, fracture) or blue sclera at birth, the majority are born fracture-free with a normal exam. In the absence of a family history, bluish sclera may be considered a normal age-related variant commonly found until 6–9 months. Teeth are generally normal and long bone fractures may not occur until children begin to walk.
OI TYPE IV OI type IV results from heterozygous missense mutations in either COL1A1 or COL1A2. De novo mutations are
308
frequent. [Patel et al., 2015] Many infants with OI type IV are fracture free at birth and have normal sclerae. Children with OI type IV may be small for their age and develop long bone bowing and fractures in childhood, but during the first two years of life, even in a family with OI, genetic testing might be required to confirm the diagnosis.
OI TYPE V OI type V results from a single mutation in IFITM5 that results in the addition of five amino acids at the amino terminal end of the encoded protein, BRIL (Bone-restricted interferon-induced transmembrane protein-like protein, which, however, is not interferon induced). The natural history of OI type V is characterized by fractures, short stature, normal sclerae, reduced bone density, hypertrophic callus and bony hypertrophy, of the radius and ulna. In a child without a family history, the diagnosis might be missed because of the age-depended, and variable expression of the disorder [Shapiro et al., 2013]. There is often a mesh-like pattern of lamellation in bone when examined under polarized light microscopy [Rauch et al., 2013]. The periosteal hyperostosis and new bone formation is often described as mineralization of the interosseous membrane between the ulna and radius and can be seen by 6 months of age [Arundel et al., 2011]. The exact mechanism of disease and the explanation for variability of the phenotype are under investigation. Until the full IFITM5 OI type V phenotypic spectrum is clear, sequencing of the gene is necessary in testing [Grover et al., 2013; Guillen-Navarro et al., 2014].
RARE RECESSIVE FORMS OF OI There are rare recessive OI phenotypes that might be overlooked in the absence of a thorough genetic evaluation OI type VI, which results from biallelic mutations in SERPINF1, often presents with few if any signs of OI at birth and the infants begin to develop progressive
309
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
scoliosis and bone deformity by about a year of age. There are few clinical clues to this diagnosis until bone deformity develops and so parental consanguinity may be the trigger for specific diagnostic studies. Bruck syndrome is a genetically
OI type VI, which results from biallelic mutations in SERPINF1, often presents with few if any signs of OI at birth, and the infants begin to develop progressive scoliosis and bone deformity by about a year of age. There are few clinical clues to this diagnosis until bone deformity develops and so parental consanguinity may be the trigger for specific diagnostic studies. heterogeneous recessively inherited disorder that results from biallelic mutations in PLOD2 or in FKBP10 [Alanay et al., 2010; Ha-Vinh et al., 2004; Schwarze et al., 2013]. Most infants with these conditions have congenital contractures, but in their absence fractures can raise concern about NAI. In this context genetic diagnosis is the usual pathway to diagnosis.
CLINICAL SIGNS THAT MAY HELP TO DISTINGUISH CHILDREN WHO HAVE BEEN ABUSED FROM THOSE WITH MILD OI The work-up for suspected physical abuse includes physical examination, radiological assessmen, and laboratory studies [Flaherty et al., 2014]. In some cases, children may have findings that can be seen with both child abuse and underlying genetic conditions. In other cases, children may have injuries that are highly suspicious for abuse, such as
bruising in the pattern of an object or hand [Anderst et al., 2013]. Posterior rib fractures, transverse fractures of several digits, metacarpals or metatarsals, and metaphyseal chip fractures were considered strongly suggestive of NAI when Kempe made his initial observations [Kempe et al., 1984]. Although each of these types of fracture can be seen in infants and children with OI, their frequency remains substantially greater in the context of trauma [Greeley et al., 2013; Shelmerdine et al., 2015]. Given the clinical importance of the distinction
Posterior rib fractures, transverse fractures of several digits, metacarpals or metatarsals, and metaphyseal chip fractures were considered strongly suggestive of NAI when Kempe made his initial observations . Although each of these types of fracture can be seen in infants and children with OI, their frequency remains substantially greater in the context of trauma. between abuse and biological causes of fracture, the recognition that no fracture type is pathognomonic of NAI has led to the search for biological tests that could distinguish between the two options.
HISTORY OF GENETIC TESTING FOR OI In the early 1970s, it was found that cultured dermal fibroblasts from individuals with different forms of OI usually made less type I procollagen than cells from controls or the chains of type I procollagen had altered electrophoretic mobilities. These tests were developed to facilitate the diagnosis of OI and to try to differentiate among individuals with different clinical presentations. Very quickly this type of
REVIEW ARTICLE
study began to be used to try to exclude the diagnosis of OI in the context of fractures thought to be related to NAI. This led Wenstrup and colleagues to ask if this test could identify abnormalities in cells from all individuals with well documented OI [Wenstrup et al., 1990]. They found that among those with non-lethal forms of OI, cells from about 87% of these individuals either made too little collagen or the collagen they made behaved abnormally when separated on polyacrylamide gels. This finding led defense attorneys to argue in criminal cases that a negative result did not exclude the diagnosis of OI “beyond a reasonable doubt.” This argument assumed that the a priori probability that a child who was tested had OI was very high, something that was rarely correct. In 1996, Steiner et al. [1996] reviewed studies of collagen production by cultured fibroblasts from 48 children who were tested to distinguish OI from NAI. Of these, six had abnormal studies consistent with OI (12.5%), but only one patient (2% of the total) had not been suspected to have OI, based on clinical findings, by the time the tests were completed. Marlowe et al., in 2002 examined the results of the same type of studies in an additional 262 children with a differential diagnosis that included NAI and OI [Marlowe et al., 2002]. They found 11 (4.1%) with a type I collagen protein abnormality consistent with OI. This study confirmed that clinical examination alone did not identify some children with mild forms of OI [Marlowe et al., 2002]. Review of clinical data from 138 children in that study, for whom sufficient clinical information accompanied the samples submitted to the CDL, revealed that in six instances the diagnosis of OI was suspected on clinical grounds by the time studies were completed. Still, reviewing all data from all the children identified in that study to have OI by this type of testing, no combination of additional clinical signs (for example, blue sclerae, short stature), associated history (family history or history of injury), or radiological features (nature and location of the fractures or Wormian
REVIEW ARTICLE
bones or bowed bones), identified all affected infants. Overall, the incidence of OI identified by analysis of collagen production by cultured fibroblasts among children evaluated for NAI was 5%, and those missed based on clinical or radiological features was about 2%.
Overall, the incidence of OI identified by analysis of collagen production by cultured fibroblasts among children evaluated for NAI was 5%, and those missed based on clinical or radiological features was about 2%. Pathogenic alterations in COL1A1 or COL1A2 account for 90–95% of individuals with non-lethal forms of OI sequenced in the CDL and about the same reported in the Database of OI Mutations (http://www.le.ac.uk/ genetics/collagen/; CDL Data). A subset of pathogenic variants that lead to changes in the amino terminal regions of the proa chains of type I procollagen escape recognition by protein screening alone [Cabral et al., 2005; Marini et al., 2007]. Because DNA-based testing more accurately identifies individuals with mild OI, analysis of collagens produced by cultured fibroblasts has taken a back seat to molecular genetic testing. Presently, genomic DNA sequence analysis of both COL1A1 and COL1A2 is the first step in laboratory testing when the diagnosis of OI is considered in most patients.
HOW DOES DNA SEQUENCE ANALYSIS COMPARE WITH PROTEIN STUDIES FOR DETECTION OF OI IN THE OI V NAI SETTING? In a recent review of molecular test requests submitted to the CDL from 2008 to 2014, there were 305 samples in
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
which NAI was noted on the request for DNA sequencing of COL1A1 and COL1A2. Ninety percent (n ¼ 276) of the children were under the age of 1 year (1–60 months) with a mean age of testing of 7 months (median 5 months). Nine children (3% of studied samples) had a pathogenic variant in COL1A1 (6) or COL1A2 (3), a frequency similar to the reported frequency in protein based testing. Two of the nine infants were also suspected to have OI by clinical evaluation at the time of testing. Thus, 7/9 infants found to have OI (2% of the total sample) were missed by clinical evaluation alone. Review of limited clinical details confirmed that those identified with OI could not be clinically distinguished (on paper) from those with normal test results. One important issue in this evaluation is that the cohort tested is a highly selected subset of all those infants and children evaluated for NAI. In most instances, the criteria used to select children to be tested remain elusive. A concern with molecular testing is that a variant of unknown significance (VUS) could be found; in most settings, when the a priori likelihood to find a pathogenic variant is low, the chance to find a VUS is the same as in the general population. A VUS was identified in 20 (6.5%) of 305 children undergoing COL1A1 and COL1A2 sequencing in this setting. Of these 12 were identified in COL1A1 and eight in COL1A2. When a VUS is identified, the usual request is to study parents (or other family members) to determine if the variant segregates with the phenotype in question. This request may lead to determination that the variant was inherited from an individual with no bone phenotype and strengthen the case against the parents—a form of selfincrimination. In this setting the ethical, legal, and social issues may prove challenging.
RECOMMENDATION FOR LABORATORY TESTING IN OI V NAI SETTING Laboratory guidelines for the diagnosis of OI were developed by the European
310
Molecular Genetics Quality Network (EMQN) on the basis of the experience of 17 OI expert clinicians and diagnostic laboratory directors [van et al., 2012]. In the setting of suspected NAI in the absence of clear phenotypic signs of OI or a family history, using EMQN guidance, testing would truncate after type I collagen gene sequence analysis was completed (See Fig. 1 from van et al., 2012). The suggested guidelines in the OI v NAI setting include analysis of IFITM5 sequence concurrently.
SUGGESTED APPROACH TO TESTING FOR OI IN THE CONTEXT OF CONCERNS ABOUT NAI At this time there are no consensus guidelines that identify the circumstances in which genetic testing for OI is indicated or when clinical and radiological evaluation alone can exclude the diagnosis of OI. Additional research on number of fractures, timing of fractures, and types of fractures seen among patients presenting with OI could help to establish definitive guidelines for molecular testing. Testing for OI can be completed as follows: 1. Sequence COL1A1, COL1A2, and
IFITM5 simultaneously. Exclude deletion or duplication of part or all of these three genes by high resolution array CGH or MLPH duplication/deletion testing. If a causative variant is not identified, in the absence of a significant pathologic clinical phenotype, no additional gene testing is indicated. Sixteen (16) genes are presently known to play a role in fracture susceptibility in children with recessively inherited forms of OI [Byers and Pyott, 2012; Cho et al., 2012; Lapunzina et al., 2010; Martınez-Glez et al., 2012; Semler et al., 2012; Shaheen et al., 2012; Pyott et al., 2013; Symoens et al., 2013; Garbes et al., 2015; Mendoza-Londono et al., 2015; Rauch et al., 2015]. In total, these account for less than 10% of all individuals identified with OI. With rare exceptions, the infants with recessively inherited forms of OI
311
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
Figure 1.
REVIEW ARTICLE
Recommended approach to genetic testing for osteogenesis imperfecta in a child with signs of non-accidental injury.
usually present with prenatal or neonatal onset of fractures, and moderate to severe bony phenotype. Second tier testing for recessively inherited forms of OI is not recommended in instances of unexplained fractures with or without blue sclera unless other features, such as consanguinity or congenital contractures, demand reconsideration. 2. If a variant of uncertain significance (VUS) is identified, complete recommended segregation studies (see discussion above). The likelihood to find a VUS in the absence of a known pathogenic variant in the type I collagens genes is 6.5% (CDL data). The VUS may be a synonymous change, an alteration in intron sequence or result in a non-glycine substitution in the triple helical
domain. The non-glycine substitutions may occur in the X or Y-positions of the triple helical domain that is characterized by the repeating Gly-Xaa-Yaa triplet that extends for 1,014 residues. Segregation studies are the first step in following up the uncertain test outcome. Of the 20 VUS reported to an ordering physician in the 2008–2014 CDL review, there were eight instances of follow-up testing of one or both parents (40%) an indication that testing recommended for final interpretation is often ignored, misunderstood, or not undertaken because of concern for the implications of such studies, as noted above. There are several obstacles to parental testing (perceived costs, parental availability, and removal of child from the family) that may also limit further tests.
Clinical evaluation for child abuse includes consideration of differential diagnoses such as accidental trauma, and underlying bone disease. In addition to a careful medical history and physical examination, the American Academy of Pediatrics report discusses laboratory, and radiology testing to assess for occult injury and bone disease [Flaherty et al., 2014]. Testing for occult injury can include liver function studies, skeletal survey and head imaging with decisions based on clinical presentation and patient age. Genetic tests, like those for OI, are targeted specifically and the diagnostic and exclusion value depend on test sensitivity and specificity, and the a priori likelihood that a child in the test environment could have OI. For example, if the a priori likelihood to have OI is 2% (based on chance of mild OI
REVIEW ARTICLE
without any clinical features or family history in NAI cases) and the test identifies 95% of people with OI, then an infant with a normal test has an approximately 1/980 chance to have OI as a consequence of a pathogenic variant in a type I collagen gene. Historically,
If the a priori likelihood to have OI is 2% (based on chance of mild OI without any clinical features or family history in NAI cases) and the test identifies 95% of people with OI, then an infant with a normal test has an approximately 1/980 chance to have OI as a consequence of a pathogenic variant in a type I collagen gene. this type of calculation has been avoided in laboratory results because of less certainty about test sensitivity, and the recognition that OI is not the only alternative explanation to NAI.
CONCLUSION As we pointed out in the Introduction, if all children with non-lethal OI under the age of 3 years had fractures that led to evaluation of NAI, they would represent a very small subset of the children evaluated. There are many tools to facilitate the diagnosis of OI and these include evaluation of family history, careful consideration of the context of fracture, an informed clinical examination, and radiographic study of the fracture and other bones. Even with these tools, evaluation of a small subset (about 5% as per our laboratory data) of the children in whom a diagnosis of NAI is considered identified biochemical or genetic alterations consistent with OI. Given the high social cost of a missed diagnosis in these children, the current practice of testing appears to be justified.
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
One of the major issues in the current selection of children to be tested is the criteria used. The number of children tested represents a very small subset of those considered each year to be at risk for NAI. It is clear that for some children the clinical tools are sufficient to make the diagnosis and testing is deferred for a later time when genetic counseling for the family, or evaluation of reproductive choices for the child might be considered. Careful evaluation of this decision process could evolve more powerful clinical mechanisms that would identify more children with OI. Currently, recommendations include DNA determination of the sequence of COL1A1, COL1A2, and IFITM5 simultaneously and exclude deletion or duplication of part or all of these three genes (array CGH or MLPH duplication/deletion testing). If a causative variant is not identified, in the absence of a significant pathologic clinical phenotype, no additional gene testing is indicated. If a VUS is found, parental segregation studies are recommended but may be hard to obtain due to follow-up or other social issues. Future studies on implications of VUS in NAI cases may offer guidance to clinicians when this situation arises. Meanwhile, in these complex cases, the molecular laboratory serves to provide a piece of puzzle. Interpreting results and implications for each child, and considering the diagnosis of NAI versus OI, ultimately involves a number of experts; evaluations are multidisciplinary and may involve medical providers, child protective services, and in some cases law enforcement professionals.
REFERENCES Alanay Y, Avaygan H, Camacho N, Utine G, Boduroglu K, Aktas D, Alikasifoglu M, Tuncbilek E, Orhan D, Bakar F, Zabel B, Superti-Furga A, Bruckner-Tuderman L, Curry C, Pyott S, Byers P, Eyre D, Baldridge D, Lee B, Merrill A, Davis E, Cohn D, Akarsu N, Krakow D. 2010. Mutations in the gene encoding the RER protein FKBP65 cause autosomal-recessive osteogenesis imperfecta. Am J Hum Genet 86:551–559. Anderst JD, Carpenter SL, Abshire TC. 2013. Section on Hematology/Oncology and
312
committee on child abuse and neglect of the American Academy of Pediatrics. Pediatrics 131:e1314–e1322. Arundel P, Offiah A, Bishop NJ. 2011. Evolution of the radiographic appearance of the metaphyses over the first year of life in type V osteogenesis imperfecta: Clues to pathogenesis. J Bone Miner Res 26: 894–898. Bronicki L, Stevenson RE, Spranger J. 2015. Beyond osteogenesis imperfecta: Causes of fractures in childhood. AJMG. Byers PH, Pyott SM. 2012. Recessively inherited forms of osteogenesis imperfecta. Annu Rev Genet 46:475–497. Cabral WA, Makareeva E, Colige A, Letocha AD, Ty JM, Yeowell HN, Pals G, Leikin S, Marini JC. 2005. Mutations near amino end of alpha1(I) collagen cause combined osteogenesis imperfecta/Ehlers-Danlos syndrome by interference with N-propeptide processing. J Biol Chem 280:19259–19269. Cho TJ, Lee KE, Lee SK, Song SJ, Kim KJ, Jeon D, Lee G, Kim HN, Lee HR, Eom HH, Lee ZH, Kim OH, Park WY, Park SS, Ikegawa S, Yoo WJ, Choi IH, Kim JW. 2012. A single recurrent mutation in the 50 -UTR of IFITM5 causes osteogenesis imperfecta type V. Am J Hum Genet 91:343–348. Flaherty EG, Perez-Rossello JM, Levine MA, Hennrikus WL, America acamady of pediatric committee on child abuse and neglect, section radiology, section on endocrinology, section on orthopaedics and the society for pediatric radiology. 2014. Evaluating children with fractures for child physical abuse. Pediatrics 133:e477–e489. Garbes L, Kim K, Rieß A, Hoyer-Kuhn H, Beleggia F, Bevot A, Kim MJ, Huh YH, Kweon HS, Savarirayan R, Amor D, Kakadia PM, Lindig T, Kagan KO, Becker J, Boyadjiev SA, Wollnik B, Semler O, Bohlander SK, Kim J, Netzer C. 2015. Mutations in SEC24D, encoding a component of the COPII machinery, cause a syndromic form of osteogenesis imperfecta. Am J Hum Genet 96:432–439. Greeley CS, Donaruma-Kwoh M, Vettimattam M, Lobo C, Williard C, Mazur L. 2013. Fractures at diagnosis in infants and children with osteogenesis imperfecta. J Pediatr Orthop 33:32–36. Grover M, Campeau PM, Lietman CD, Lu JT, Gibbs RA, Schlesinger AE, Lee BH. 2013. Osteogenesis imperfecta without features of type V caused by a mutation in the IFITM5 gene. J Bone Miner Res 28:2333–2337. Guillen-Navarro E, Ballesta-Martınez MJ, Valencia M, Bueno AM, Martinez-Glez V, L opezGonzalez V, Burnyte B, Utkus A, Lapunzina P, Ruiz-Perez VL. 2014. Two mutations in IFITM5 causing distinct forms of osteogenesis imperfecta. Am J Med Genet A 164A:1136–1142. Ha-Vinh R, Bank AY, Campos-Xavier RA, Zankl AB, Superti-Furga A, Bonafe A. 2004. Phenotypic and molecular characterization of Bruck syndrome (osteogenesis imperfecta with contractures of the large joints) caused by a recessive mutation in PLOD2. Am J Med Genet A 131:115–120. Kempe CH, Silverman FN, Steele BF, Droegemueller W, Silver HK. 1962. Landmark article July 7, 1962: The battered-child
313
AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMINARS IN MEDICAL GENETICS)
syndrome. By C. Henry Kempe, Frederic N. Silverman, Brandt F. Steele, William Droegemueller, and Henry K. Silver. JAMA 251:3288–3294. Lapunzina P, Aglan M, Temtamy S, Caparr osMartın J, Valencia M, Let on R, MartınezGlez V, Elhossini R, Amr K, Vilaboa N, Ruiz-Perez V. 2010. Identification of a frameshift mutation in osterix in a patient with recessive osteogenesis imperfecta. Am J Hum Genet 87:110–114. Leventhal JM, Martin KD, Asnes AG. 2010. Fractures and traumatic brain injuries: Abuse versus accidents in a US database of hospitalized children. Pediatrics 126: e104–e115. Marini JC, Forlino A, Cabral WA, Barnes AM, San Antonio JD, Milgrom S, Hyland JC, K€ orkk€ o J, Prockop DJ, De Paepe A, Coucke P, Symoens S, Glorieux FH, Roughley PJ, Lund AM, Kuurila-Svahn K, Hartikka H, Cohn DH, Krakow D, Mottes M, Schwarze U, Chen D, Yang K, Kuslich C, Troendle J, Dalgleish R, Byers PH. 2007. Consortium for osteogenesis imperfecta mutations in the helical domain of type I collagen: Regions rich in lethal mutations align with collagen binding sites for integrins and proteoglycans. Hum Mutat 28:209–221.Review. Marlowe A, Pepin M, Byers P. 2002. Testing for osteogenesis imperfecta in cases of suspected non-accidental injury. J Med Genet 39:382–386. Martınez-Glez V, Valencia M, Caparr os-Martın JA, Aglan M, Temtamy S, Tenorio J, Pulido V, Lindert U, Rohrbach M, Eyre D, Giunta C, Lapunzina P, Ruiz-Perez VL. 2012. Identification of a mutation causing deficient BMP1/mTLD proteolytic activity in autosomal recessive osteogenesis imperfecta. Hum Mutat 33:343–350. Mendoza-Londono R, Fahiminiya S, Majewski J, Tetreault M, Nadaf J, Kannu P, Sochett E, Howard A, Stimec J, Dupuis L, Roschger P, Klaushofer K, Palomo T, Ouellet J, Al-Jallad H, Mort JS, Moffatt P, Boudko S, B€achinger HP, Rauch F, Consortium CRC. 2015. Recessive osteogenesis imperfecta caused by missense mutations in SPARC. Am J Hum Genet 96:979–985.
Patel RM, Nagamani SC, Cuthbertson D, Campeau PM, Krischer JP, Shapiro JR, Steiner RD, Smith PA, Bober MB, Byers PH, Pepin M, Durigova M, Glorieux FH, Rauch F, Lee BH, Hart T, Sutton VR. 2015. A cross-sectional multicenter study of osteogenesis imperfecta in North Americaresults from the linked clinical research centers. Clin Genet 87:133–140. Pyott SM, Tran TT, Leistritz DF, Pepin MG, Mendelsohn NJ, Temme RT, Fernandez BA, Elsayed SM, Elsobky E, Verma I, Nair S, Turner EH, Smith JD, Jarvik GP, Byers PH. 2013. WNT1 Mutations in families affected by moderately severe and progressive recessive osteogenesis imperfecta. Am J Hum Genet 92:590–597. Rauch F, Fahiminiya S, Majewski J, Carrot-Zhang J, Boudko S, Glorieux F, Mort JS, B€achinger HP, Moffatt P. 2015. Cole-Carpenter syndrome is caused by a heterozygous missense mutation in P4HB. Am J Hum Genet 96:425–431. Rauch F, Moffatt P, Cheung M, Roughley P, Lalic L, Lund AM, Ramirez N, Fahiminiya S, Majewski J, Glorieux FH. 2013. Osteogenesis imperfecta type V: Marked phenotypic variability despite the presence of the IFITM5 c.-14C>T mutation in all patients. J Med Genet 50:21–24. Schwarze U, Cundy T, Pyott SM, Christiansen HE, Hegde MR, Bank RA, Pals G, Ankala A, Conneely K, Seaver L, Yandow SM, Raney E, Babovic-Vuksanovic D, Stoler J, Ben-Neriah Z, Segel R, Lieberman S, Siderius L, Al-Aqeel A, Hannibal M, Hudgins L, McPherson E, Clemens M, Sussman MD, Steiner RD, Mahan J, Smith R, Anyane-Yeboa K, Wynn J, Chong K, Uster T, Aftimos S, Sutton VR, Davis EC, Kim LS, Weis MA, Eyre D, Byers PH. 2013. Mutations in FKBP10, which result in Bruck syndrome and recessive forms of osteogenesis imperfecta, inhibit the hydroxylation of telopeptide lysines in bone collagen. Hum Mol Genet 22:1–17. Semler O, Garbes L, Keupp K, Swan D, Zimmermann K, Becker J, Iden S, Wirth B, Eysel P, Koerber F, Schoenau E, Bohlander SK, Wollnik B, Netzer C. 2012.
REVIEW ARTICLE A mutation in the 5’-UTR of IFITM5 creates an in-frame start codon and causes autosomal-dominant osteogenesis imperfecta type V with hyperplastic callus. Am J Hum Genet 91:349–357. Shaheen R, Alazami AM, Alshammari MJ, Faqeih E, Alhashmi N, Mousa N, Alsinani A, Ansari S, Alzahrani F, Al-Owain M, Alzayed ZS, Alkuraya FS. 2012. Study of autosomal recessive osteogenesis imperfecta in Arabia reveals a novel locus defined by TMEM38B mutation. J Med Genet 49:630–635. Shapiro JR, Lietman C, Grover M, Lu JT, Nagamani SC, Dawson BC, Baldridge DM, Bainbridge MN, Cohn DH, Blazo M, Roberts TT, Brennen FS, Wu Y, Gibbs RA, Melvin P, Campeau PM, Lee BH. 2013. Phenotypic variability of osteogenesis imperfecta type V caused by an IFITM5 mutation. J Bone Miner Res 28:1523– 1530. Shelmerdine SC, Arthurs OJ, Calder AD. 2015. Rib fractures in osteogenesis imperfecta: Have we learnt anything about child abuse? J Pediatr Orthop 35:e81. Sillence D, Senn A, Danks D. 1979. Genetic heterogeneity in osteogenesis imperfecta. J Med Genet 16:101–116. Stoll C, Dott B, Roth M, Alembik Y. 1989. Birth prevalence rates of skeletal dysplasias. Clin Genet 35:88–92. Symoens S, Malfait F, D’hondt S, Callewaert B, Dheedene A, Steyaert W, B€achinger HP, De Paepe A, Kayserili H, Coucke PJ. 2013. Deficiency for the ER-stress transducer OASIS causes severe recessive osteogenesis imperfecta in humans. Orphanet J Rare Dis 8:154. van Dijk FS, Byers PH, Dalgleish R, Malfait F, Maugeri A, Rohrbach M, Symoens S, Sistermans EA, Pals G. 2012. EMQN best practice guidelines for the laboratory diagnosis of osteogenesis imperfecta. Eur J Hum Genet 20:11–19. Wenstrup R, Willing M, Starman B, Byers P. 1990. Distinct biochemical phenotypes predict clinical severity in nonlethal variants of osteogenesis imperfecta. Am J Hum Genet 46:975–982.
