Osteogenesis Imperfecta

[Brittle Bone Disease, OI]

Clinical Diagnosis

The clinical diagnosis of OI depends on the presence of a number of features (see Table 1). No "diagnostic criteria," such as those for the types of Ehlers-Danlos syndrome, exist for types of OI. Features of OI:

• Fractures with minimal or no trauma in the absence of other factors, such as abuse or other known disorders of bone

• Short stature or stature shorter than predicted based on stature of unaffected family members, often with bone deformity

• Blue sclerae

• Dentinogenesis imperfecta (DI)

• Progressive, post-pubertal hearing loss

• Additional clinical features including ligamentous laxity and other signs of connective tissue abnormality

• Family history of OI, usually consistent with autosomal dominant inheritance

Radiographic features of OI (Table 2). The radiographic features of OI change with age. The major findings include the following:

• Fractures of varying ages and stages of healing, often of the long bones but may also include ribs and skull. The metaphyseal chip fractures characteristic of abuse can be seen in a small number of children with OI.

• "Codfish" vertebrae, which are the consequence of spinal compression fractures, seen primarily in the adult

• Wormian bones, defined as "sutural bones which are 6 mm by 4 mm (in diameter) or larger, in excess of 10 in number, with a tendency to arrangement in a mosaic pattern" [Cremin et al 1982]. Wormian bones are suggestive of, but not pathognomic for, OI.

• Protrusio acetabuli (The socket of the hip joint is too deep and the acetabulum bulges into the cavity of the pelvis causing intrapelvic protrusion of the acetabulum.)

• Osteopenia detected by dual energy X-ray absorptiometry (DEXA). Of note, bone density can be normal, especially in OI type I, as DEXA measures mineral content rather than collagen content [Deodhar & Woolf 1994, Paterson & Mole 1994, Lund et al 1999]. In OI, changes with age of bone mineral density of the spine are primarily caused by increases in vertebral bone volume and secondarily caused by increased mineral density [Kurtz et al 1985, Moore et al 1998, Reinus et al 1998].

  Note: (1) A major determinant of bone density may be the patient's ability to ambulate. (2) Bone density standards for children under the age of two years have been determined after sampling very small populations, often fewer then ten, so that reliability is an issue. (3) Bone density standards for children are based on height; corrections for short stature of severely affected individuals may need to be made.

Testing

Histomorphometric evaluation of iliac crest bone from individuals with types V, VI, and VII OI shows maintained lamellar structure, reduced cortical width and cancellous bone volume, and increased bone remodeling [Rauch et al 2000]. The distinction of OI type VI from OI types IV and V rests in part on a characteristic "fish scale" appearance under polarized light.

Analysis of type 1 collagen synthesized in vitro by culturing dermal fibroblasts obtained from a small skin biopsy shows the structure and quantity of the collagen. The sensitivity of biochemical testing is 87% in non-lethal forms of OI and about 98% in the lethal form. The sensitivity varies with clinical phenotype and reflects the small effect of some COL1A1 or COL1A2 mutations on type I collagen structure and quantity [Wenstrup et al 1990] as well as absence of COL1A1 or COL1A2 mutations in OI types V, VI, and VII, which were not recognized at the time of the survey completed by Wenstrup et al (1990).

Molecular Genetic Testing

GeneReviews designates a molecular genetic test as clinically available only if the test is listed in the GeneTests Laboratory Directory by either a US CLIA-licensed laboratory or a non-US clinical laboratory. GeneTests does not verify laboratory-submitted information or warrant any aspect of a laboratory's licensure or performance. Clinicians must communicate directly with the laboratories to verify information.—ED.

Genes. COL1A1 and COL1A2 are the two genes associated with OI types I, II, III, and IV. They encode the two chains pro α1(I) and pro α2(I), respectively, of type I procollagen.

Other loci

• OI type VII maps to a 4.5-Mb region on the short arm of chromosome 3.

• The loci for OI types V and VI have not been mapped.

Molecular genetic testing: Clinical uses

• Diagnostic confirmation

• Prenatal diagnosis

Molecular genetic testing: Clinical methods

Sequence analysis. Sequence analysis of COL1A1 and COL1A2 cDNA to detect mutations in the coding sequence and sequence analysis of COL1A1 and COL1A2 genomic DNA to detect mutations that alter either sequence or stability of mRNA identify close to 100% of mutations in the type I collagen genes. The mutations in most families are unique; only a few recurrent mutations (mostly CpG dinucleotides) are seen in more than one family. Mutation detection rate varies by OI type (Table 3).

The RNA for cDNA sequencing is generally derived from dermal fibroblasts. Although RNA from circulating leukocytes has been used, concerns about mono-allelic amplfication remain because of the very small amounts of illegitimate transcripts available.

Genomic DNA can be isolated from any tissue of the individual being tested.

Table 3 summarizes molecular genetic testing for this disorder.

Table 3. Molecular Genetic Testing Used in Osteogenesis Imperfecta

Test Method	Mutations Detected 1	Mutation Detection Rate	Test Availability
cDNA sequence analysis	Missense mutations, small insertions or deletions, exon-skipping mutations of COL1A1 and COL1A2	OI type I: ~100% OI type II: 98% OI type III: 60-70% 2 OI type IV: 70-80% 2	Clinical
Genomic DNA sequence analysis	Nonsense mutations, missense mutations, splice-site mutations of COL1A1 and COL1A2

1. Description of mutations in type I collagen genes follows the usual conventions with some exceptions (see Molecular Genetic Pathologenesis).
2. In OI type III, the frequency of identified mutations in type I collagen genes is lower than expected for dominant disorders; rare instances of recessive inheritance caused by mutations in type I collagen genes have been identified. Other genes no doubt also play a role.

Interpretation of test results

• For issues to consider in interpretation of sequence analysis results, click here.

• It is often important to have mRNA derived from cultured cells to interpret "sequence variants" that have not been previously recognized.

• Analysis of genomic DNA may miss instances of whole gene deletion unless attention is paid to the state of common polymorphic nucleotides.

• Sequence variants or small deletions that underlie the primers used for amplification of genomic DNA can lead to mono-allelic amplification; thus small deletions will be missed.

• Although some features of the methods used may result in unsuccessful detection of alterations in collagen quantity or structure and/or mutation in COL1A1 or COL1A2, failure to identify such abnormalities substantially reduces the likelihood that the tested individual has OI types I, II, III, or IV.

• In the case of some mutations — exon skipping or small deletions — it may be important to characterize the mutation itself at the genomic DNA level to be sure of the precise mutation.

• It may be important to characterize the effects of some genomic mutations to be sure that they account for an abnormality at the mRNA or protein level.

Testing Strategy for a Proband

The sensitivity for indentifying OI by molecular genetic testing is similar to the sensitivity of analysis of the structure and quantity of type 1 collagen in cultured fibroblasts from a skin biopsy.

Genetically Related Disorders

Classic Ehlers-Danlos syndrome (EDS). Classic Ehlers-Danlos syndrome, which includes EDS type I and EDS type II, is characterized by skin hyperextensibility, abnormal wound healing, and joint hypermobility. Bones are not unusually fragile and fractures are not increased. Classic EDS is inherited in an autosomal dominant manner. More than 50% of individuals with classic EDS have an identifiable mutation in either the COL5A1 gene or the COL5A2 gene. A single nucleotide alteration that results in substitution of arginine by cysteine at position 134 (R134C) of the triple helix of the pro α1(I) chains encoded by one COL1A1 allele has been described in two unrelated individuals with classic EDS [Nuytinck et al 2000]. The arginine residue is highly conserved in fibrillar collagens and localized to the X position of the Gly-X-Y triplet; how this mutation is related to the phenotype is not known.

Ehlers-Danlos Syndrome (EDS) arthrochalasia type, (formerly EDS VIIA and VIIB). Affected individuals have congenital bilateral hip dislocation, short stature, joint hypermobility, osteopenia, kyphoscoliosis, velvety hyperextensible skin, and mild bone fragility. The mutations that cause EDS type VIIA (in COL1A1) and VIIB (in COL1A2) affect the splicing of exon 6, which encodes the site at which the amino-terminal procollagen protease cleaves the amino-terminal propeptide from procollagen in the course of extracellular processing. Because all or part of the exon is missing from the final product, the pro αchains retain the extended propeptide, which interferes with normal fibrillogenesis.

Osteoporosis. Polymorphisms in COL1A1 or COL1A2, particularly a variant in the COL1A1 first intron, have been associated with osteoporosis [Grant et al 1996, Uitterlinden et al 1998].

Arterial dissection. Mayer et al (1996) described a G-to-C transversion in one COL1A1 allele resulting in a G13A substitution in the triple helical domain of the α1(I) chain of type I collagen in a 35-year-old woman with dissection of the right internal carotid artery and the right vertebral artery after scuba diving. Other than a history of easy bruising and bluish sclerae, she had no evidence of a connective tissue disorder or of OI. Family history was negative for vasculopathy.

Natural History

The severity of OI ranges from perinatal lethality to individuals with severe skeletal deformities, mobility impairments, and very short stature to nearly asymptomatic individuals with a mild predisposition to fractures, normal stature, and normal life span. Before the molecular basis of OI was understood, OI was classified into four types based on clinical presentation, radiographic features, family history, and natural history [Sillence et al 1979]. Recent classifications based on additional studies that include bone histomorphometry and molecular genetic characterization have identified additional types of OI (Table 1). OI types V, VI, and VII were identified among those with "OI type IV" because of unique clinical signs, abnormalities in bone, and the absence of COL1A1 and COL1A2 mutations; it is unclear if they will be grouped with OI in the future.

Although this classification into types of OI is helpful in providing information about prognosis and management, the features of different types of OI overlap and it is not always easy to classify individuals. It is helpful to remember that the severity of clinical and radiographic features lies on a continuum and that the "types" are defined using characteristics that appear to form clinical "nodes." Interfamilial variability is apparent among individuals with the same OI type and intrafamilial variability is apparent among individuals with the same mutation. Nonetheless, it is reasonable to continue to think of OI in terms of these types in order to provide information about the expected clinical course.

OI type I. OI type I is characterized by blue sclerae and normal stature. A small proportion of infants with OI type I have femoral bowing at birth. The first fractures may occur at birth or with diapering. More often, the first fractures occur when the infant begins to walk and, more importantly, to fall. Fractures generally occur at a rate of a few to several per year and then decrease in frequency after puberty. Fracture frequency often increases again late in adulthood, especially in postmenopausal women and men beyond the fifth decade [Paterson et al 1984]. Affected individuals may have anywhere from a few fractures to more than 100, but the fractures usually heal normally with no resulting deformity.

Most affected individuals have normal or near normal stature, but may be short compared to other members of their families.

Joint hypermobility may contribute to morbidity.

In their classification of OI, Sillence et al (1979) designated a subset of OI type I with DI (OI type IB). In DI, morbidity results not from dental decay, but rather from premature wearing down of the teeth, which interferes with chewing. DI can be a significant cosmetic issue. Dental eruption in OI type I can sometimes occur early.

Progressive hearing loss occurs in about 50% of adults with OI type I, beginning as a conductive hearing loss but becoming a sensorineural hearing loss in time.

OI type II. Abnormalities characteristic of OI type II are evident at birth. Weight and length are small for gestational age. The sclerae are dark blue and connective tissue is extremely fragile. The skull is large for the body size and soft to palpation. Callus formation on the ribs may be palpable. Extremities are short and bowed. Hips are usually flexed and abducted in a "frog-leg" position. Although some fetuses with OI type II die in utero or are spontaneously aborted, more typically infants die in the immediate perinatal period. More than 60% of affected infants die on the first day; 80% die within the first week; survival beyond one year is exceedingly rare and usually involves intensive support such as continuous assisted ventilation [Byers et al 1988]. Death usually results from pulmonary insufficiency related to the small thorax, rib fractures, or flail chest because of lack of stable ribs. Those who survive the first few days of life may not be able to take in sufficient calories because of respiratory distress.

Histologic evaluation of bone from infants with OI type II shows marked reduction in collagen in secondary trabeculae and cortical bone [Horton et al 1980]. Cortical bone is hypercellular with large osteocytes. Trabeculae contain woven bone with large immature osteoblasts [Cole et al 1992, Cole & Dalgleish 1995].

OI type III. The diagnosis of OI type III is readily apparent at birth. Fractures in the newborn period, simply with handling of the infant, are common. In some affected infants, the number and severity of rib fractures leads to death from pulmonary failure in the first few weeks or months of life. Infants who survive this period generally fare well, although most do not walk without assistance and usually use a wheel chair or other assistance for mobility because of severe bone fragility and marked bone deformity. Affected individuals have as many as 200 fractures and progressive deformity even in the absence of fracture. OI type III is difficult to manage, even with intramedullary rod placement. Growth is extremely slow and adults with OI type III are among the shortest individuals known, with some having adult stature of less than one meter (three feet). Intellect is normal unless there have been intracerebral hemorrhages.

Even within OI type III, there is considerable heterogeneity at the clinical level. Some individuals have normal-appearing teeth and facial appearance while others have DI and a large head with enlarged ventricles that reflect the soft calvarium. Relative macrocephaly and barrel chest deformity are observed. Usually sclerae are blue in infancy but lighten with age. Hearing loss generally begins in the teenage years.

Basilar impression, an abnormality of the craniovertebral junction caused by descent of the skull on the cervical spine, is common. Basilar impression is characterized by invagination of the margins of the foramen magnum upward into the skull, resulting in protrusion of the odontoid process into the foramen magnum. Basilar impression may progress to brainstem compression, obstructive hydrocephalus, or syringomyelia because of direct mechanical blockage of normal CSF flow [Charnas & Marini 1993, Sillence 1994, Hayes et al 1999].

Symptoms become apparent with neck flexion. Findings include posterior skull pain, C2 sensory deficit, tingling in the 4th and 5th digits, and numbness in the medial forearm. When swimming, affected individuals may perceive that water temperature differs below and above the umbilicus. Lhermitte's sign (tingling on neck flexion) can be demonstrated at any stage. Basilar impression can cause headache with coughing, trigeminal neuralgia, loss of function of the extremities, or parasthesias. At its most severe levels of involvement, sleep apnea and death can occur.

OI type IV. OI type IV is characterized by mild short stature, DI, adult-onset hearing loss, and normal-to-grey sclerae. This is the most variable form of OI, ranging in severity from moderately severe to so mild that it may be difficult to make the diagnosis. Stature is variable and may vary markedly within the family. DI is common but may be mild. Sclerae are typically light blue or gray at birth but quickly lighten to near normal. Hearing loss occurs in some. Basilar impression/invagination can occur.

OI type V. This group was initially thought to have either OI type III or OI type IV because of short stature and fractures. Sclerae were generally white. Two features distinguish OI type V: striking hypertrophic callus formation, usually at the site of fractures and most often in the femoral shaft, and calcification of the interosseous membrane between the ulna and the radius that leads to inability to fully supinate and pronate the forearm.

OI type VI. The clinical features of OI type VI are similar to those of OI type IV. The defining features of OI type V are absent.

OI type VII. OI type VII is distinguished by rhizomelic shortening of all limbs.

Genotype-Phenotype Correlations

Genotype-phenotype correlation is not readily available and prediction of phenotype and natural history for unknown mutations has no precise rules.

OI type I almost always results from mutations in one COL1A1 allele that introduce premature termination codons and decrease the stability of mRNA. These mutations may occur by codon changes, by frame shifts, and by splice mutations that result in use of cryptic splice sites and frame shifts. The type I collagen molecule contains two pro α1(I) chains and a single α2(I) chain. If the number of available pro α1(I) chains decreases, the amount of the trimer manufactured is diminished because no more than one pro α2(I) chain can be accommodated per molecule.

Individuals with OI type IB (i.e., OI type I with DI) have mutations that alter the sequence of type I collagen chains, while those without DI have mutations that change the amount of type I collagen made.

OI types II, III, and IV all result from mutations that alter the structure of either pro α1(I) or pro α2(I) chains. The most common mutations result in substitution of another amino acid for glycine in the triple helical domain of either chain; serine, arginine, cysteine, and tryptophan result from substitutions in the first position of the glycine codon and alanine, valine, glutamic acid, and aspartic acid result from substitutions in the second position of the glycine codon. Substitutions in the pro α1(I) chain by arginine, valine, glutamic acid, aspartic acid, and tryptophan are almost always lethal if they occur in the carboxyl-terminal 70% of the triple helix and have a non-lethal, but still moderately severe phenotype, if they occur in the remainder of the chain. For the smaller side-chain residues (serine, alanine, and cysteine), the phenotypes are more variable and appear to reflect some characteristics of the stability profile of the triple helix that are not yet fully recognized. Much more variability occurs with mutations that affect glycine residues in the pro α2(I) chain, even with the large side-chain residues; therefore, it is more difficult to determine the genotype-phenotype relationship.

The other common mutations affect splice sites. Mutations that lead to exon skipping in the pro α1(I) chain beyond exon 14 and in the pro α2(I) chain beyond exon 25 are generally lethal. The phenotypes resulting from mutations in the upstream region are more variable and may lead to significant joint hypermobility.

A relatively small number of mutations that alter amino acid sequences in the carboxyl-terminal regions of both chains has been identified. These domains are used for chain association and mutations have the capacity to destroy this property or lead to abnormalities in chain association. The phenotypic effects of mutations that affect this domain appear to be milder when they result in exclusion rather than inclusion of the chain.

Mosaicism for dominant mutations has been recognized in OI types II, III, and IV. The phenotype of the mosaic individual can range from no identifiable characteristics of OI to one of the mild forms. Individuals mosaic for mutations that result in non-lethal forms of OI generally have no phenotypic features of OI, even when the mutation is present in a majority of somatic cells. Mosaicism for mutations that result in lethal OI can produce a mild OI phenotype if the mutation is present in the majority of somatic cells; otherwise, the mosaicism is generally asymptomatic.

Penetrance

The penetrance of the dominantly inherited forms of OI for which affected individuals are heterozygous for a COL1A1 or COL1A2 mutation is 100%, although expression may vary considerably, even in the same family.

Nomenclature

The classification scheme of "OI congenita" and "OI tarda" was discarded because fractures at birth can be noted in mild OI and infants with severe OI may not have fractures at birth.

In classifications of genetic conditions, OI may be considered a skeletal dysplasia, a connective tissue disorder, a disorder of collagen or extracellular matrix, or a disorder of bone fragility.

Prevalence

Considering all types, OI has a prevalence of approximately 6-7/100,000. The two mildest forms, OI type I and OI type IV, account for considerably more than half of all OI. OI is found in all racial and ethnic groups. A study in Edinburgh revealed an incidence of those types detectable at birth of one in 20,000 [Wynne-Davies & Gormley 1981]. The most recent estimates of prevalence come from Finland, where OI is thought to affect about 1/15,000 individuals [Kuurila et al 2002]. If the same prevalence is assumed for the US, then about 18,000 people in the US would be affected.

• OI type I has a prevalence of approximately 3-4/100,000. Sillence (1979) reported an incidence in Australia of 3.5 per 100,000.

• OI type II has an incidence of about 1-2/100,000. [Sillence (1979) reported an incidence in Australia of 1.6 per 100,000], translating to an approximate prevalence of 1/20,000 in those who survive early childhood mortality. OI type II is often not considered in prevalence data because of early lethality. Increasingly, OI type II is recognized in utero as a consequence of early screening by ultrasound examination, so that the birth frequency in developed countries is dropping.

• OI type III has a prevalence of 1-2/100,000. Sillence (1979) reported an incidence in Australia of 1.6 per 100,000 for OI type III.

• Sillence (1979) believed OI type IV to be uncommon. Since then it has been recognized that this is a relatively common form of OI, probably about as common as OI type I.

• OI type V and OI type VI each account for about 5% of people with OI.

• OI type VII has been found to date only in a Native Canadian population.

In Utero

Early prenatal ultrasound examination or radiographic findings may lead to a consideration of hypophosphatasia, thanatophoric dysplasia, campomelic dysplasia, and achondrogenesis as well as OI type II. Experienced sonographers usually have little difficulty in distinguishing among these disorders. In some cases, either biochemical or molecular testing can be a useful adjunct.

Infancy and Childhood

• Bruck syndrome (OMIM 259450) [Viljoen et al 1989, McPherson & Clemens 1997] is an autosomal recessive condition characterized by bone fragility, congenital joint contractures, clubfeet, normal or blue sclerae, and wormian bones. It results from defects in the lysyl hydroxylase that hydroxylates the amino-terminal lysyl residues involved in crosslink formation [Bank et al 1999].

• Osteoporosis pseudoglioma syndrome (OMIM 259770) includes bone fragility and fractures, other skeletal deformities, pseudoglioma with blindness in infancy, and other anomalies. It is caused by mutations in the gene encoding the lipoprotein receptor-related protein 5 [Gong et al 1996, 2001].

• Cole-Carpenter syndrome (OMIM 112240) is characterized by bone deformities, multiple fractures, ocular proptosis, shallow orbits, orbital craniosynostosis, frontal bossing, and hydrocephalus [Cole & Carpenter 1987].

• Hadju-Cheney syndrome (OMIM 102500) is characterized by short stature, failure to thrive, conductive hearing loss, dysmorphic features, early tooth loss, genitourinary anomalies, osteopenia, pathologic fractures, wormian bones, failure of suture ossification, basilar impression, vertebral abnormalities, kyphoscoliosis, cervical instability, joint laxity, dislocation of the radial head, long bowed fibulae, pseudoclubbing, short distal digits, acroosteolysis, and hirsutism.

• Gerodermia osteodysplastica (OMIM 231070) is characterized by dwarfism, lax skin, osteoporosis, wormian bones, fractures, vertebral compression, dysmorphic facies.

• Idiopathic juvenile osteoporosis (IJO) typically presents in pre-adolescents with fractures and osteoporosis. The fracture susceptibility and osteoporosis usually resolve spontaneously with puberty. The etiology of IJO is unknown.

• Dentiogenesis imperfecta. DI can occur separately from OI as an isolated familial condition as a result of mutations in the DSPP gene on chromosome 4 [Rajpar et al 2002].

• Non-accidental trauma (child abuse). OI needs to be distinguished from child physical abuse/non-accidental trauma. The prevalence of physical abuse is much greater than the prevalence of OI. On rare occasion, both can occur in the same child. Patient history, family history, physical examination, radiographic imaging, and the clinical course all contribute to the distinction of OI from child abuse. The overlap in clinical features includes multiple or recurrent fractures, fractures that do not match the history of trauma, and the finding of fractures of varying ages and at different stages of healing [Carty 1988, Ablin et al 1990, Steiner et al 1996, Ablin & Sane 1997, Marlowe et al 2002]. The continued occurrence of fractures in a child who has been removed from a possibly abusive situation lends support to the possibility of OI. Metaphyseal and rib fractures, thought to be virtually pathognomonic for child abuse, can occur in OI.

  
  The presence or absence of blue sclerae is unreliable in distinguishing OI from child abuse because blue sclerae are often found in unaffected normal infants untilabout 18 months of age; children with OI type IV may not have blue sclerae.

  
  Family history is often unrevealing; families suspected of possible child abuse often provide an unverified family history of frequent fractures; conversely, the family history of individuals with OI often does not reveal any other affected individuals because of a de novo mutation in the proband or the presence of a mild phenotype in relatives.

  
  Laboratory testing (protein-based studies or molecular genetic testing of COL1A1 and COL1A2) often is not needed, and in some cases the time required to perform such testing can delay proper disposition of child abuse cases.

Evaluations at Initial Diagnosis

• Hearing evaluation

Treatment of Manifestations

Management focuses on supportive therapy to minimize fractures and maximize function, minimize disability, foster independence, and maintain overall health [Marini & Gerber 1997]. Ideally, OI is managed by a multidisciplinary team including specialists in medical therapy of OI, orthopedics, and rehabilitation medicine. Supportive therapy is individualized depending upon the severity, the degree of impairment, and the age of the patient. Considerable support is generally required by medical personnel to help parents feel comfortable caring for infants with OI type II.

Physical medicine treatment

• Instruction of parents and other caregivers in safe handling techniques

• Bracing of limbs with such devices as vacuum pants that are rigid

• Orthotics to stabilize lax joints

• Promotion of appropriate physical activity

• Physical and occupational therapy for increased stability of bone, improved mobility, prevention of contractures, prevention of head and spinal deformity, aerobic fitness, and muscle strengthening

• Endurance training [Marini 1988, Bleck 1981, Gerber et al 1990, Binder et al 1993, Marini & Gerber 1997]

• Mobility devices such as scooters and chairs for children and modified automobiles for adults

• Analgesics for pain

Orthopedic treatment

• Fractures are treated as they would be in unaffected children and adults with attention to the following:

  • The period of immobility in children with OI should be shortened as much as is practical.

  • Casts should be small and lightweight.

  • Physical therapy should begin as soon as the cast is removed to promote mobility and enhance muscle strength and bone mass.

  • At this time, intramedullary rodding remains a mainstay of orthopedic care to provide anatomic positioning of limbs that permits more normal function. It is not yet clear if treatment with bisphosphonates will change the natural history of bone deformity and diminish the use of surgical intervention (see discussion of bisphosphonate treatment below).

• Progressive spinal deformities are particularly difficult because of the poor quality of bone in severely affected children. Progressive scoliosis in severe OI does not respond to conservative management and response to surgical intervention may be limited.

Basilar impression. Basilar impression is best evaluated by CT or MRI scanning with views across the base of the skull. Criteria for surgical intervention are not well defined. If surgery is undertaken, it should be done in a center experienced in the procedures used.

Dental treatment. The goals are the maintenance of both primary and permanent dentition, a functional bite or occlusion, optimal gingival health, and overall appearance. Pediatric dentists are the most knowledgeable about DI in children. Some consensus exists that early dental restorative coverage of the primary molars, and if possible, aesthetic coverage of the upper anterior teeth is optimal. Plastic polymers are sometimes used to coat teeth. As anxiety can be an issue with children, pre-medication for anxiolysis, such as nitrous oxide analgesia or midazolam, can be used for treatment in a clinic setting.

If warranted, orthodontic treatment can be initiated, but care must be taken not to fracture teeth with orthodontic appliances because of the brittleness of the teeth.

Dental restorations in adults may best be done by a general dentist knowledgeable about OI or a specialist in prosthetic dentistry.

Treatment of hearing loss. Initial hearing loss in OI is usually conductive as a result of fractures of the bones of the middle ear or contracture and scarring of the incus. Surgical repair of the bones and creation of a prosthetic incus can improve unaided hearing.

Later hearing loss appears to have a significant sensorineural component that does not respond to middle ear surgery. Cochlear implantation has been used in a small number of individuals; outcome data are limited.

Prevention of Secondary Complications

General anesthesia. Special attention should be paid to anesthesia concerns including proper positioning on the operating room table, for which egg crate foam is recommended.

Surveillance

Children with OI should be seen by the dentist twice yearly, beginning in early childhood or even infancy.

Hearing evaluation should be performed at 3-5 year intervals after adolescence until hearing loss is identified.

Therapies Under Investigation

Bisphosphonates, analogs of pyrophosphate that decrease bone resorption, are being evaluated in both uncontrolled and controlled trials to assess the extent to which they can increase bone mass and bone strength and improve function in children with OI. These studies are still ongoing. Bisphosphonates have been used most extensively in severely affected children with OI; they may be useful in adults with OI as well [Adami et al 2003]. Because of the lack of published randomized controlled clinical trials, biosphosphonates should used with caution in the treatment of OI outside of research trials.

An open-label trial of cyclical intravenous Pamidronate (bisphosphonate) was reported by Glorieux and colleagues (1998) and the effects of relatively long term use in adults by Astrom and Soderhall (2002) and Zeitlin et al (2003). Falk et al (2003) replicated the study of Glorieux et al (1998) in children over 22 months of age, but did report one child with fracture non-union following treatment with Pamidronate. No randomized, placebo-controlled clinical trial of Pamidronate has been published. Bisphosphonate treatment has produced improvements in bone histomorphometry, increased bone mineral density (BMD), decreased some biochemical markers of bone resorption, and possibly reduced fracture risk. In addition, some investigators have reported decreased bone pain in young, but not older, children following treatment.

Pamidronate use is invasive and inconvenient, typically requiring intravenous infusions every three months four hours a day for three days and has real and potential complications. Recently, Pamidronate has been offered even to very young children with OI, but complications including transient asymptomatic hypocalcemia [Plotkin et al 2000] and symptomatic hypocalcemia [Chien et al 2002] have been noted. The long-term consequences of lowering bone turnover in children with OI are unknown, but may include delayed bone union after fracture or osteotomy.

A randomized, controlled clinical trial using the oral bisphosphonate, Alendronate, was recently completed and the results are being analyzed.

A newer bisphosphonate, Zoledronic acid, with a longer half life, greater potency, and more convenient dosing is being studied in children with OI.

Injection of human growth hormone has been used in clinical trials in children with OI and appears to improve linear growth rates and bone formation in some children [Antoniazzi et al 1996].

Bone marrow transplantation (BMT) to introduce normal mesenchymal stem cells that have the capacity to differentiate into normal osteoblasts is being evaluated [Horwitz et al 1999, 2001].

Search ClinicalTrials.gov for access to information on clinical studies for a wide range of diseases and conditions.

Other

Management of pregnancy and delivery. Pregnancy presents two issues: The risk to the mother who has OI and the risk to the child of having OI (See Genetic Counseling section). Delivery by cesarian section and by vagina have about the same rate of complications for infants with each type of OI. Delivery by cesarian section is more frequent than in the general population because a non-vertex presentation cannot be corrected by external manipulation. Women with significant skeletal deformity and short stature should be followed by high-risk prenatal care centers.

Management of lethal OI. It is appropriate to offer the parents the option of allowing the infant to expire without attempting interventions such as assisted ventilation.

Other therapies. Early trials of anabolic steroids, sodium fluoride, testosterone, vitamins C and D, flavinoids, and calcitonin showed minimal or no improvement in bone formation, or too small a sample size was utilized for meaningful conclusions [reviewed in Byers & Steiner 1992].

Genetics clinics are a source of information for individuals and families regarding the natural history, treatment, mode of inheritance, and genetic risks to other family members as well as information about available consumer-oriented resources. See the GeneTests Clinic Directory.

Support groups have been established for individuals and families to provide information, support, and contact with other affected individuals. The Resources section may include disease-specific and/or umbrella support organizations.

Mode of Inheritance

Osteogenesis imperfecta types I, II, IV, and V are inherited in an autosomal dominant manner. Osteogenesis imperfecta type III is usually inherited in an autosomal dominant manner; rare instances of autosomal recessive inheritance have been observed. OI type VII is inherited in an autosomal recessive manner. The mode of inheritance of OI type VI is not yet certain.

Risk to Family Members - Autosomal Dominant

Parents of a proband

• Many individuals diagnosed with the milder forms of OI (type I and some probands with type IV) have an affected parent.

• The proportion of cases caused by de novo mutations varies by the severity of disease. Approximately 60% of individuals with mild OI have de novo mutations; virtually 100% of individuals with severe (type III) and lethal (type II) OI have de novo mutations.

Recommendations for the evaluation of parents of a proband with an apparent de novo mutation include clinical examination of the parents and molecular genetic testing if the mutation in the proband has been identified.

Sibs of a proband

• If a parent of the proband is affected, the risk to the sibs is 50%.

• When the parents are clinically unaffected, the risk to the sibs of a proband is about 5% because somatic and/or germline mosaicism in a parent is possible.

Offspring of a proband. Each child of an individual with a dominantly inherited form of OI has a 50% chance of inheriting the mutation.

Other family members. The risk to other family members depends upon the status of the proband's parents. If a parent is found to be affected, his or her family members are at risk.

Risk to Family Members - Autosomal Recessive

Parents of a proband

• The parents of an affected child are obligate heterozygotes and therefore carry one mutant allele.

• Heterozygotes (carriers) are asymptomatic.

Sibs of a proband

• At conception, the sibs of an affected individual have a 25% chance of being affected, a 50% chance of being asymptomatic carriers, and a 25% chance of being unaffected and not carriers.

• Once an at-risk sib is known to be unaffected, the risk of his/her being a carrier is 2/3.

• Heterozygotes (carriers) are asymptomatic.

Offspring of a proband. The offspring of an individual with OI type VII are obligate heterozygotes (carriers) for a disease-causing mutation.

Other family members of a proband. Sibs of the proband's parents are at 50 % risk of being carriers.

Carrier Detection

Carrier testing using molecular genetic techniques is not offered because it is not clinically available.

Related Genetic Counseling Issues

Considerations in families with an apparent de novo mutation. When neither parent of a proband with an autosomal dominant condition has clinical evidence of the disorder, it is likely that the proband has a de novo mutation or that one parent has germline mosaicism with or without somatic mosaicism. Other explanations, including alternate paternity or undisclosed adoption, could also be explored.

Family planning. The optimal time for determination of genetic risk and discussion of the availability of prenatal testing is before pregnancy.

DNA banking. DNA banking is the storage of DNA (typically extracted from white blood cells) for possible future use. Because it is likely that testing methodology and our understanding of genes, mutations, and diseases will improve in the future, consideration should be given to banking DNA of affected individuals. DNA banking is particularly relevant in situations in which the sensitivity of currently available testing is less than 100%. See DNA Banking for a list of laboratories offering this service.

Prenatal Testing

High-risk pregnancy

• Molecular genetic testing. Analysis of DNA extracted from fetal cells obtained by amniocentesis usually performed at about 15-18 weeks' gestation* or chorionic villus sampling (CVS) at about 10-12 weeks' gestation is possible if the disease-causing allele of an affected family member has been identified.

• Biochemical analysis of collagen from fetal cells obtained by CVS at about 10-12 weeks' gestation has been reported. An abnormality of collagen from cultured cells of the proband must be identified before this technique can be used for prenatal testing. Biochemical analysis of collagen from amniocytes is not useful because amniocytes do not produce type I collagen.

Note: OI type I cannot be identified prenatally even with CVS cells because the proportion of type I procollagen produced by the cells is reduced compared to control cells and resemble cells from individuals with OI. Thus false positive testing can be an issue.

• Prenatal ultrasound examination performed in a center with experience in diagnosing OI, done at the appropriate gestational age, can be a valuable tool in the prenatal diagnosis of OI. Normally, ultrasound examination detects only the lethal and most severe forms of OI prior to 20 weeks' gestation; fetuses affected with milder forms may be detected later in pregnancy when fractures or deformity occur.

  • OI type II: The bony abnormalities can first be seen by ultrasound examination by about 13-14 weeks' gestation. By 16 weeks, femoral length is typically two or more weeks delayed, calvarial mineralization is essentially absent, and ribs generally have identified fractures.

  • OI type III: Limb length generally begins to fall off the growth curve at about 17-18 weeks' gestation; serial ultrasound examinations are required.

Low-risk pregnancy. Routine prenatal ultrasound examination may identify a fetus not known to be at risk for OI with findings suggestive of OI (Type II or III) including bowed, crumpled femurs, beaded ribs, evidence of fractures, and markedly diminished calvarial mineralization. As a part of the evaluation of such findings, molecular genetic testing for COL1A1 or COL1A2 mutations may be considered. Inability to identify a mutation, however, does not eliminate the diagnosis of OI in the fetus.

*Gestational age is expressed as menstrual weeks calculated either from the first day of the last normal menstrual period or by ultrasound measurements.

Molecular Genetic Pathogenesis

More than 90% of individuals with OI have mutations in one of the two genes, COL1A1 and COL1A2, that encode the chains of type I procollagen, the major protein in bone and most other connective tissues. The two general outcomes of these mutations are either a decrease in the amount of type I procollagen produced or the production of some abnormal type I procollagen molecules. OI type I is generally characterized by decreased production of type I procollagen. In the vast majority of instances, OI type I results from mutations in the COL1A1 gene that result in premature termination codons. The majority of these mutations are deletions or insertions of a small number of nucleotides, a number not divisible by three, in the coding sequences of exons throughout the gene. These mutations, single codon changes that introduce premature termination codons, and some splice-site mutations that lead to exclusive use of cryptic sites and generation of out-of-frame transcripts all lead to premature termination codons. The presence of a premature termination codon that is separated by one or more introns in the gene leads to marked instability of the mRNA derived from the mutant allele. As a consequence, the amount of COL1A1 mRNA is reduced to half the normal amount, with no compensation by the other allele. Type I procollagen is a trimer that must contain at least two pro α1(I) chains. With a reduction in the COL1A1 mRNA, an obligatory decrease in the production of type I procollagen occurs, although the protein produced is structurally normal. The diminished amount of type I collagen in bone appears to reduce the amount of bone that can be made and leads to brittle bones.

The majority of mutations that result in OI type II, III, and IV result in substitutions for glycine within the triple helical domain of the pro αchain. In the products of both genes, the triple helical domain contains 1014 amino acids in the repeating sequence of Gly-X-Y without interruption. The glycine residues are essential for the normal folding of the molecule and any substituting amino acid (eight possibilities result from single nucleotide substitutions in glycine codons) appears to result in OI. These mutations are designated by the position of glycine (1-1012) in the triple helical domain. There are 160 amino acids amino terminal to this position in the pro α1(I) chain and 90 in the pro α2(I) chain. In the major databases, both positions are generally reported.

The other types of OI result from synthesis of pro αchains that have altered sequences. The vast majority of these mutations result in substitutions of glycine residues in the triple helical portion of the chains of type I procollagen. The pro αchains consist of an amino-terminal propeptide, a triple helical segment of 1014 amino acids in which glycine is in every third position and prolines preceding glycine residues are generally hydroxylated as are some lysyl residues in the Y-position of the Gly-X-Y triplet. Glycine, the smallest amino acid, must be in the third position to allow proper chain folding to occur. When substituted, the propagation of the triple helix is delayed, additional post-translation modification occurs, and some of the assembled trimers are never secreted. The consequence of these mutations is that a diminished amount of type I procollagen is secreted and some of the protein in the matrix has an abnormal structure. Mutations in the two genes result in a different proportion of abnormal molecules, but all phenotypes can result from mutations in either gene. The clinical consequence in these circumstances appears to result from the position of the substituted glycine, the chain in which the substitution occurs, and the nature of the substituting amino acid.

Similar phenotypes (OI type II, III, or IV) can result from short deletions or duplications of single amino acids or Gly-X-Y triplets and from exon-skipping events. The relationship between genotype and phenotype is complex, although it appears that mutations closer to the 5' end of the coding sequence that affect the amino-terminal ends of the triple helical domains are likely to have milder clinical phenotypes. This is probably the reflection of chain association occurring at the carboxyl-terminal end of the chain so that a small part of the molecule is affected. Much is still to be learned about the final pathways of molecular pathogenesis.

Literature Cited

Published Statements and Policies Regarding Genetic Testing

No specific guidelines regarding genetic testing for this disorder have been developed.