Statistics from Altmetric.com
Congenital kyphomelia, or bowing of the femora, is associated with a number of skeletal dysplasias that include campomelic dysplasia, Stüve-Wiedemann dysplasia, Bruck syndrome, Antley-Bixler syndrome, Fuhrmann syndrome, and osteogenesis imperfecta (OI).12 In most reported cases, the femora become progressively more angulated with age. However, spontaneous resolution of congenitally bowed femora has been recognised to occur in a small number of cases associated with either OI or a rare skeletal dysplasia known as kyphomelic dysplasia (KD).3–5
Osteogenesis imperfecta is a connective tissue disorder that is caused in more than 90% of cases by an abnormality of type I collagen. Clinical manifestations of OI may include bone fragility and/or deformities, blue sclerae, short stature, joint laxity, deafness, Wormian bones, and dental abnormalities. Owing to considerable phenotypic variability in OI, a classification system based on clinical, genetic, and radiographic characteristics has been used for the last 20 years to divide this diagnostic category into four broad clinical subtypes.6 While all types of OI may present with congenital bowing of the long bones, especially the femora, this finding is most commonly associated with types II and III, the neonatal lethal and progressively deforming types, respectively.
In 1983, Maclean et al described an infant with broad, angulated femora and several minor skeletal abnormalities that included a narrow thorax, platyspondyly, and micrognathia.7 An unusual feature of the proband's skeletal deformity was that the bowing improved considerably over the first six months of life. The pattern of skeletal involvement and the atypical natural history were felt to represent a novel skeletal dysplasia that was named “kyphomelic dysplasia” (OMIM 211350). Over the subsequent two decades, at least 15 cases with phenotypic similarities to kyphomelic dysplasia have been reported.8 Other phenotypic findings noted in some cases of apparent KD include dimpling of the skin overlying long bone deformities, variable bowing of other long bones, rhizomelic shortening, metaphyseal irregularities, a small thorax often with 11 flared ribs, and platyspondyly.9–11 The diagnosis of KD has been periodically challenged since its inception. Pitt12 considered a case of apparent KD to represent a variant of femoral hypoplasia-unusual facies syndrome. Cisarik et al13 described four patients with KD with widely variable manifestations including a “classical lethal” form and questioned whether the more severe phenotype could represent either an allelic variant or a distinct entity. More recently, the existence of KD was further challenged following revision of the diagnosis of the index case described by Maclean et al to Schwartz-Jampel syndrome.14
This report presents the clinical course and molecular analysis of the type I collagen genes of a 35 year old woman initially thought to have KD, but whose molecular studies have shown that she has a variant of OI. The histories of her two affected children are also presented to provide insight into the prenatal and neonatal manifestations of this disorder.
The proband is a 35 year old woman who presented at 23 weeks' gestation following a routine fetal ultrasound that showed generalised rhizomelic and mesomelic shortening of the limbs with severe angulation malformations of the femora (fig 1). In reviewing her past medical history, the proband reported that she was born with severely bowed legs. Review of films from infancy showed that she had rhizomesomelic shortening of the lower extremities with severely bowed femora and fibulae, and Wormian bones in the calvarium (fig 2A). At the age of 3 months, she was diagnosed with osteogenesis imperfecta after a reportedly normal evaluation for hypophosphatasia. From approximately 3 months to 1 year of age, her legs were maintained in full length plaster casts, which slightly improved the bowing (fig 2B). After starting to walk, her lower limbs showed dramatic improvement over the subsequent year (fig 2C). The proband also has a history of significant dental decay, particularly involving the primary teeth. A recent dental evaluation has shown that her dental abnormalities are consistent with dentinogenesis imperfecta. She bruises easily and has broken three bones in her life time, each after significant trauma. Over the past few years, she has developed worsening lumbar spondylolisthesis. Her intellectual development was normal and she has neither hearing loss nor joint laxity. There is no previous family history of short stature, blue sclerae, poor dentition, osseous fragility, easy bruising, hearing loss, or joint laxity.
At presentation, she had a height of 147 cm (below the 5th centile) with a weight and head circumference measuring within the 50th-75th centiles. Her upper to lower segment ratio and arm span to height ratio were on the 95th centile. She had pale blue sclerae and teeth with translucent enamel and multiple caries. Her limbs were normally proportioned without joint laxity or limitation. Her skin had normal elasticity and there were no abnormal scars. A radiological survey of her skeleton showed femora with very mild varus deformities (fig 2D). Other abnormalities noted in the skeletal survey included a malunion fracture of the left clavicle, grade 3 spondylolisthesis involving the L5/S1 vertebral bodies, and Wormian bones in the calvarium. She had no radiographic signs of osteopenia or anomalies of the scapulae or vertebrae.
The proband's pregnancy was followed with serial ultrasonographic evaluations at 28 and 32 weeks' gestation. The fetal femora showed worsening of bowing during gestation with lengths following established growth curves,15 but remaining below the 3rd centile. In contrast, bilateral tibial and unilateral radial bowing noted in the initial sonogram improved during gestation. The final sonogram also noted slightly decreased echogenicity of the calvarium.
At 39 weeks' gestation, a male infant was delivered by caesarean section because of a history of a previous low transverse caesarean section. The infant weighed 2685 g (between the 5th and 10th centiles), measured 43.5 cm in length (below the 5th centile, median height of 33 weeks' gestation), had a crown-rump length of 33.5 cm (50th centile), and a head circumference of 34.5 cm (between the 25th and 50th centiles). His sclerae had a faint bluish hue and his face had no appreciable dysmorphic features. The chest was normally shaped with a circumference measuring between the 25th and 50th centiles. The only anomalies noted on physical examination were severe varus deformities of both thighs without dimpling of the overlying skin. Radiographic skeletal survey showed femoral bowing with slight metaphyseal flaring, Wormian bones, and the presence of only 11 ribs bilaterally (fig 3). The infant had an uncomplicated course during the neonatal period and was healthy during early infancy. However, he died unexpectedly at 6 weeks of age with the cause of death being attributed to sudden infant death syndrome (SIDS). Necropsy showed abnormal rib morphology with anterior flaring and evidence of a healing left femoral fracture. Histological analysis of samples from the left femur, skull, and ribs showed osteopenia and focal areas of marrow fibrosis.
In a subsequent pregnancy, prenatal ultrasonography at 18 weeks' gestation identified a fractured right femur, bowed right tibia, fractured right humerus, and shortened and bowed left femur. A follow up sonogram at 34 weeks' gestation showed worsening of the bowing of the femora and tibiae. Following a term pregnancy, a female was born by caesarean section with a weight and head circumference falling within the 25th-50th centiles whereas the length of 40.5 cm fell below the 5th centile. Physical examination at birth was notable for grey sclerae, rhizomelic shortening of all extremities, and bowing of the femora and tibiae. There was cutaneous dimpling over the lateral aspects of the midshafts of the femora and humeri. A radiographic survey of the skeleton showed diffuse osteopenia, several enlarged costochondral junctions consistent with healing fractures, and bilateral shortening and bowing of the femora and tibiae. A radius to humerus ratio of 0.91 confirmed rhizomelic shortening of the upper extremities according to standard curves.16
A skin biopsy was obtained from the forearm of the proband with appropriate consent, and fibroblasts were grown by routine methods. The collagens and procollagens synthesised by these cells were analysed as previously described.17 The cells were found to produce both a normal and an abnormal species of type I procollagen (fig 4A, B). In the latter, the chains were overmodified along the amino-terminal 40% of the triple helical domain, consistent with a mutation in either the COL1A1 or COL1A2 gene. Exons 6-52 of the COL1A1 and COL1A2 genes were amplified by PCR and screened for mutations by conformation sensitive gel electrophoresis.18 Several heteroduplexes were identified and the sequence of the exon and flanking splice junctions were determined for each. The only pathogenic abnormality identified was a G to C transversion (1406G→C) with the A of the initiator methionine as the nucleotide at position 1 of the mRNA, CAGGGCCTC→ CAGGCCCTC) in one allele (fig 4C). This sequence alteration is predicted to result in substitution of the glycine at position 379 of the triple helical domain of the proα2(I) chain by alanine (G379A, in which the first glycine of the triple helix is referred to as the index position or G469A when the initiator methionine is the reference point). Karyotypic analysis of the male first child showed a normal 46,XY genotype.
The mild skeletal abnormalities, short stature, and dentinogenesis imperfecta combined with the molecular studies confirm that the proband and her children have osteogenesis imperfecta type IVB. The persistent pale blue sclera is uncommon in OI type IV, but can be seen in up to 10% of affected subjects. The striking improvement of the congenital deformities, to our knowledge, has not been reported in OI type IV. Improvement of femoral bowing in OI has only been described for type I, and there is little documentation of the natural history of the improvement in these cases.3,19 An interesting aspect of the improvement in femoral bowing noted in this report is that it coincided with the start of walking, which may suggest that weight bearing plays a role in the remodelling process.
Before the type I collagen studies, the diagnoses of osteogenesis imperfecta and kyphomelic dysplasia were both considered. Initially, the diagnosis of KD seemed more likely owing to the natural history of the proband's femoral bowing and phenotypic features in her infants, such as cutaneous dimpling over bowed femora, metaphyseal irregularities, and 11 pairs of ribs with flaring at the costochondral junctions.8 However, the identification of a type I collagen mutation established the diagnosis of osteogenesis imperfecta type IV for these affected family members. This report provides further evidence that at least some of the cases of apparent kyphomelic dysplasia are actually mild or atypical forms of other skeletal dysplasias. To our knowledge, no other cases of presumed KD have been examined molecularly. As other cases carrying the diagnosis of KD are analysed for mutations in either the type I collagen genes or Perlecan, the gene recently shown to be mutated in cases of Schwartz-Jampel syndrome,20 it should be possible to determine whether KD exists as a specific diagnostic entity.
Type I collagen is a triple helix composed of two chains encoded by the COL1A1 gene and one chain encoded by the COL1A2 gene. In the more than 200 mutations identified in these two genes, there are no predominant mutations and most mutations are restricted to single families. The majority of mutations in both genes result in single amino acid substitutions for glycine residues within the triple helical domain of the chains. In this region, glycine occurs in every third position and substitution by any other amino acid, which introduces a bulkier side chain, is predicted to impair formation of the collagen triple helix.21 Of the 55 missense substitutions described in the COL1A2 gene that replace glycine residues, only two mutations have resulted in substitutions with alanine (at positions 199 which results in OI type IV22 and 1006 which results in OI type III23). Alanine codons result from substitutions for the second position G by C in all glycine codons (GGN). These substitutions are rare in both type I collagen genes and in the type III collagen gene (COL3A1), mutations which result in Ehlers-Danlos syndrome type IV phenotype.24 There appear to be two possible explanations for the low frequency of these substitutions: either the phenotypes do not correspond to those produced by other substitutions for glycine because of the low bulk of the side chain methyl group, or G→C transversions in these codons are uncommon. There are insufficient data to distinguish between these two possibilities, although substitutions for glycines at positions 91025 and 92826 in the COL1A1 gene both result in the classical lethal OI type II phenotype, and neither of these cases, to our knowledge, has features of resolving kyphomelia. Thus, mutation frequency may be the more likely explanation.
The clinical courses of the infants presented in this report raise several important points. First, the significant prenatal bowing at 18 weeks' gestation would have raised a greater degree of concern for a more severe form of OI or other skeletal dysplasia had it not been for the natural history of the mother. The lack of multiple fractures in the femora and the absence of significant demineralisation of the calvaria in the early stages of both pregnancies argued against a lethal form of a skeletal dysplasia,27 but the significant deformities with progression throughout pregnancy were felt to be consistent with a number of more severe syndromes, such as OI type III, Stüve-Wiedemann syndrome, or Schwartz-Jampel type 2 syndrome.28 This case suggests that a potentially mild clinical course can follow when angulated femora are detected in early pregnancy or at birth. Second, the early death of the male infant in this report prompted a review of infant mortality in cases of apparent KD and OI. A recent review of the majority of reported cases with presumed KD noted that four out of 12 babies died between 2 and 13 months of age with the cause of death being unclear in several cases.4 Temple et al10 described a single case of an infant with apparent KD who had documented episodes if apnoea, including one life threatening episode, between the ages of 4-6 months. In contrast, there were no cases of infant mortality in a review of causes of death in 79 patients with confirmed OI (type II excluded).29 The risk of infant mortality in variants of OI with resolving kyphomelia will only become established as more cases are identified.
This case report expands the phenotypic and molecular spectrum of OI type IV. This phenotype with its remarkable natural history is important to appreciate for clinicians who encounter fetuses or neonates with bowing of the femora. The incorporation of this phenotype into the diagnostic category of OI supports the view that congenital bowing of the femora, even if there is striking resolution during infancy, should be considered as a symptom rather than a specific disease. The expanding range of prognoses associated with congenital bowing of long bones provides an impetus for pursuing genotype-phenotype correlations in these cases.
This work was supported in part by a grant from the National Institutes of Health, AR41223.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.