Osteoporosis of Genetic Cause and Osteoporosis-Pseudoglioma Syndrome

Non-collagen monogenic primary osteoporoses of early onset: LRP5, WNT1, PLS3, SGMS2 and related entities

Literature and regulatory information updated: 29 October 2026.

0. Initial Reference Table

The table brings together the entities of the class covered in this article. Osteogenesis imperfecta (OI) due to variants in the type I collagen genes and the recessive forms of OI, as well as the high bone mass disorders caused by gain of function of LRP5 (osteopetroses and sclerosing dysplasias), appear only in the differential diagnosis.

1. Introduction

1.1 Definition of the class and of the entities

Osteoporosis is a skeletal disease characterised by low bone mass and altered microarchitecture, with increased fragility and fracture risk; it mainly affects older people, but the primary and secondary forms of childhood and young adulthood are now recognised as entities in their own right, with a true prevalence that is unknown and depends on the definition used [37]. Early-onset osteoporosis (EOOP) is osteoporosis of the child and young adult: in the growing child, it requires a clinically significant fracture history together with low bone mineral density (BMD), or a vertebral compression fracture without relevant trauma; in adults, it applies to premenopausal women and men under 50 years of age with a Z-score ≤ −2.0 or a T-score ≤ −2.5 at the lumbar spine or femoral neck, associated with fragility fractures or chronic disease [5, 32]. In the absence of known chronic disease, a targeted work-up finds a secondary cause in up to 90% of cases; when none is found, the next hypothesis is a monogenic bone disease, and the diagnosis of idiopathic osteoporosis only applies once secondary causes and known monogenic forms have been excluded [37].

This article covers, as a class, the monogenic primary osteoporoses that do not result from defects of type I collagen or of the pathways of its post-translational modification. The founding entity is osteoporosis-pseudoglioma syndrome (OPPG), an autosomal recessive disorder with severe childhood osteoporosis and congenital or infantile blindness, caused by biallelic loss of function of LRP5 [3, 27]. The same canonical WNT signalling pathway explains autosomal dominant osteoporosis due to monoallelic loss-of-function variants of LRP5 and dominant osteoporosis due to monoallelic variants of WNT1, whose biallelic counterpart is a recessive form of OI (type XV) [5, 42]. Outside the WNT pathway, X-linked osteoporosis due to PLS3 variants mainly affects hemizygous males, and bone fragility with calvarial “doughnut” lesions, due to heterozygous SGMS2 variants, combines childhood osteoporosis, sclerotic cranial lesions and cranial nerve palsies [33, 43]. Idiopathic juvenile osteoporosis (IJO) is included as a diagnosis of exclusion, with a self-limiting course in most patients [9].

In the 2023 Nosology of Genetic Skeletal Disorders, all of these entities belong to group 26, renamed from “OI and decreased bone density” to “OI and bone fragility”, because fragility is the common feature regardless of BMD; the nosology adopted dyadic names linking phenotype to gene (for example, “osteoporosis–AD form, LRP5-related”) [4]. The group also includes, as closely related entities, X-linked osteoporosis due to MBTPS2, ARHGAP25-related autosomal dominant osteoporosis, spondyloocular dysplasia due to XYLT2 and osteoporosis with developmental delay and microcephaly due to COPB2, which appear in the differential diagnosis [4].

1.2 Historical overview

OPPG was first described by Saraux in 1967, according to Streeten and colleagues [38]. In 1985, Beighton, Winship and Behari presented it as an “ocular form of osteogenesis imperfecta”, a new autosomal recessive syndrome, a name that survives as a synonym [3]. In the same year, Frontali, Stomeo and Dallapiccola, describing a sibship with an affected boy and affected monozygotic twin girls, proposed vitreoretinal dysplasia or phthisis bulbi associated with radiological osteoporosis as minimum criteria, confirmed recessive inheritance by segregation analysis and noted that the published families came more often from Mediterranean countries [44]. Superti-Furga, Steinmann and Perfumo discussed in 1986 the distinction between OPPG and OI, a distinction that remains central to the differential diagnosis [3].

In 1996, Gong and colleagues assigned OPPG to chromosomal region 11q12-13 by linkage mapping [45]. The founding article of the molecular era, published in Cell in 2001 by Gong, Warman and the OPPG Collaborative Group, identified loss-of-function variants of LRP5 as the cause of the disease, showed that heterozygous carriers have lower bone mass than age- and sex-matched controls, that osteoblasts express LRP5 in situ and that the protein transduces canonical WNT signalling; it concluded that WNT signalling through LRP5 regulates bone accrual during growth and peak bone mass [46]. In 2005, Ai and colleagues published the first large molecular series, with 37 probands [27], and Hartikka and colleagues described heterozygous LRP5 variants in children with primary osteoporosis without signs of OI [30].

The year 2013 broadened the class. Three groups identified WNT1 as an essential ligand of the pathway in bone: Laine and colleagues found a heterozygous missense variant in ten members of a Finnish family with dominant osteoporosis and a homozygous nonsense variant in a sibship with recessive OI [42]; Keupp and colleagues described five homozygous variants in consanguineous families with recessive OI and one family with dominant osteoporosis [47]; Pyott and colleagues described four families with moderately severe, progressive recessive OI [48]. In the same year, van Dijk and colleagues identified PLS3 variants in five families with X-linked osteoporosis [43]. The association between early-onset osteoporosis and sclerotic cranial “doughnut” lesions was described in 1974; the molecular cause, heterozygous SGMS2 variants, was only identified in 2019 by Pekkinen and colleagues, in six families [33, 37]. The 2018 edition of Spranger's atlas of bone dysplasias still describes the molecular basis of this syndrome as unknown [49].

Idiopathic juvenile osteoporosis was characterised by Dent and Friedman in 1965, and Smith published his experience with 21 patients in 1995, according to the review in Spranger's atlas [9].

1.3 Nomenclature and reclassifications

The term “pseudoglioma” refers to retrolental masses that, on examination, mimic an intraocular tumour and require retinoblastoma to be excluded; they correspond to fibrovascular tissue of persistent fetal vasculature and to retinal detachment [3]. OPPG and LRP5-related familial exudative vitreoretinopathy (FEVR) are now regarded by several authors as the extremes of a single loss-of-function continuum [34, 38]. In OMIM, WNT1 has two phenotype numbers: 615220 for biallelic OI type XV and 615221 for susceptibility to autosomal dominant osteoporosis [5, 12, 13]; the 2023 Nosology table assigns number 615220 to “osteoporosis–dominant form, WNT1-related”, which is a discrepancy with OMIM [4]. PLS3-related osteoporosis is recorded by OMIM as X-linked dominant and by Orphanet as X-linked recessive, and has been called “X-linked OI” by some authors [2, 7, 15]. The SGMS2 disorder carries the nosological name “bone fragility with calvarial “doughnut” lesions, SGMS2-related”, with or without spondylometaphyseal dysplasia [4, 29].

2. Molecular Genetics and Pathophysiology

2.1 The canonical WNT pathway in bone

Canonical WNT signalling regulates the skeleton from development to the maintenance of bone mass in adulthood. When a WNT ligand, such as WNT1, binds to the seven-transmembrane receptor Frizzled and to the co-receptor LRP5 (or LRP6), Dishevelled is recruited and the β-catenin destruction complex (axin, APC, GSK3 and CK1α) is inactivated; β-catenin accumulates in the cytosol, enters the nucleus and activates genes of osteoblast differentiation and function [5]. Formation of the WNT–Frizzled–LRP5 complex recruits AXIN1 to LRP5 and activates TCF/LEF-dependent transcription [27]. The physiological brake on the pathway is sclerostin, secreted mainly by osteocytes, which binds to LRP5/LRP6 and inhibits bone formation [25]; DKK1 has a similar inhibitory effect [38]. In the eye, LRP5 acts with FZD4 as a co-receptor for norrin (the NDP product), a non-WNT ligand required for retinal vascularisation; the seven OPPG missense mutants tested by Ai and colleagues reduced norrin signal transduction, consistent with the hypothesis that visual loss results from deficient norrin signalling and the bone disease from deficient WNT signalling [27].

2.2 LRP5: structure, allelic spectrum and direction of effect

LRP5 is located at 11q13.2 and encodes a 1615-amino-acid receptor of the LDL receptor superfamily, with four extracellular β-propeller (YWTD) domains interspersed with epidermal growth factor-like domains, a transmembrane domain and a cytoplasmic tail; the MANE transcript is NM_002335.4 [1, 38, 50]. The cytoband appears as 11q13.4 in Spranger's atlas and as 11q11–12 in the Oxford Desk Reference; the current location in the OMIM database is the reference one [1, 3, 51]. There is an LRP5 pseudogene on chromosome 22, which requires PCR primers for exons 1 and 3–9 to be designed so as not to amplify it [27].

The same gene produces opposite phenotypes depending on the direction of the effect. Complete, biallelic loss of function causes OPPG; partial or monoallelic loss of function is associated with autosomal dominant osteoporosis and with FEVR; gain of function causes autosomal dominant high bone mass disorders (familial high bone mass, Worth-type endosteal hyperostosis, autosomal dominant osteopetrosis type 1) [5, 52, 53]. In Spranger's atlas, osteopetrosis type A1 (OMIM 607634) due to heterozygous LRP5 variants may present at any age, from birth to 68 years, and is often asymptomatic [54, 55]. The high bone mass variants are missense variants confined to the first β-propeller domain and alter the binding site of the endogenous inhibitors (DKK1 and sclerostin, which binds to LRP5 and inhibits the WNT pathway [56]); when tested by Ai and colleagues, seven high bone mass variants did not reduce WNT signalling [27, 38]. LRP5 is also associated with isolated polycystic liver disease (OMIM 617875) and is a susceptibility factor for retinopathy of prematurity, according to Orphanet [4, 20].

In the series of Ai and colleagues, the 44 probable OPPG variants comprised 11 nonsense, 11 frameshift, 2 splicing and 20 missense variants, most in the second and third β-propeller domains, which have high affinity for WNT ligands [27, 38]. Four missense mutants (NM_002335.4:c.731C>T p.(Thr244Met), NM_002335.4:c.1067C>T p.(Ser356Leu) and, in the published protein designation, NP_002326.2:p.(Thr390Lys) and NP_002326.2:p.(Gly520Val); in the historical abbreviated notation T244M, S356L, T390K, G520V) did not transduce WNT1 or WNT10b signal; G404R and D434N had less than 50% and G610R 60% of normal activity, although the patients carrying them did not have a milder phenotype; some mutants did not reach the membrane (T390K), others were poorly trafficked [27]. No mutant interfered with the co-expressed normal receptor, which excludes a dominant-negative effect and points to functional haploinsufficiency as the mechanism of low bone mass in carriers [27]. Obligate OPPG carriers have reduced BMD without evident eye disease, an observation made in the founding article, in the Israeli family of Lev and colleagues and in the Pennsylvania Mennonites, where 4 of 6 parents had osteopenia or osteoporosis [38, 46, 57].

The limits of detection were documented early. In the 2005 series, exon sequencing found about 70% of the expected alleles; exon deletions, intronic splicing variants, amplification failures and regulatory variants escaped the strategy [27]. Laine and colleagues described splicing variants that only cDNA analysis or exon trapping confirmed: NM_002335.4:c.1584+4A>T, which abolishes the exon 7 donor site and activates a cryptic site with an in-frame insertion of 21 amino acids (p.Glu528_Val529ins21, published protein designation) and defective intracellular transport; NM_002335.4:c.4112-2A>G and NM_002335.4:c.1015+1G>T, with skipping of exons 20 and 5; and a homozygous deletion of exon 4 which, in the cDNA, removed exons 4 and 5 [58]. Narumi and colleagues identified a 7.2 kb microdeletion by high-density oligonucleotide microarray [3]. Chung and colleagues described a Turkish family with an OPPG-like phenotype, profound hypotonia and mild intellectual disability, with a deletion of 6 of the 9 consecutive leucines of the signal peptide (NM_002335.4:c.43_60del p.(Leu15_Leu20del)), which prevents entry into the endoplasmic reticulum, and a deletion of exons 14 to 16 generated by recombination between Alu repeats; the polyleucine repeat of the signal peptide is polymorphic in the general population and the different alleles alter receptor processing in vitro [59].

Some genotypes illustrate correlations. In the Mennonites, homozygosity for NM_002335.4:c.1275G>A p.(Trp425Ter) (W425X), which truncates the protein at 425 amino acids, gave classic OPPG with a mean lumbar Z-score of −4.6, close to the published mean of −4.7 ± 0.9; compound heterozygosity W425X/NM_002335.4:c.1225A>G p.(Thr409Ala) gave the same blindness, but normal stature and a mean Z-score of −1.9 (p=0.02), suggesting a missense variant with residual WNT function sufficient for bone and insufficient for the eye [38]. In Greece, a 174-base-pair deletion spanning the end of an exon and part of the following intron (NM_002335.4:c.2409_2503+79del), in homozygosity, was found in 10 patients from eight related nuclear families [60]. Alonso and colleagues showed, by three-dimensional modelling, that the variants NM_002335.4:c.235T>C p.(Trp79Arg) and NM_002335.4:c.2254C>T p.(Arg752Trp) destabilise β-propeller domains; the mother heterozygous for R752W had low BMD [61].

2.3 Osteoporosis due to monoallelic LRP5 and LRP6 variants

Hartikka and colleagues, in 20 Canadian children with primary osteoporosis without extraskeletal signs and without COL1A1/COL1A2 variants, found heterozygous LRP5 variants in three: NP_002326.2:p.(Ala29Thr) (published protein designation), NM_002335.4:c.3107G>A p.(Arg1036Gln) and a frameshift at codon 913, the latter also present in the father and brother with significant osteoporosis, and R1036Q in the mother and two osteoporotic siblings [30]. In 372 German adults with early-onset osteoporosis, after exclusion of secondary causes, Stürznickel and colleagues found relevant LRP5 or LRP6 variants in 50 individuals (31 index cases and 19 relatives), 42 in LRP5 and 8 in LRP6: 17 classified as causal, 14 as variants of uncertain significance (VUS) and 19 as BMD-associated polymorphisms; one patient had OPPG due to compound heterozygosity [28]. Collet and colleagues, in 123 young French adults, reported the variant NM_002335.4:c.1999G>A p.(Val667Met) in 22 (17.8%) and novel or very rare variants in 16 (13%), with reduced activity in a luciferase reporter assay [62]. p.Val667Met is a common variant in the European population, so its individual contribution requires caution, and the review by Mäkitie and Zillikens stresses that rare LRP5 variants classified as pathogenic, likely pathogenic or uncertain exist in the general population and are difficult to link to a specific patient without family segregation [37]. Common LRP5 variants, such as rs4988300 and rs634008, are associated with lower PINP and BMD in patients with common osteoporosis [37].

2.4 WNT1: two dose-dependent disorders

WNT1 is located at 12q13.12 and the MANE transcript is NM_005430.4 [2, 50]. Biallelic loss-of-function forms cause recessive OI (type XV, OMIM 615220), either progressively deforming (Sillence III) or moderate (Sillence IV), with structural brain malformations and rhizomelia; ptosis has been proposed as a hallmark of the recessive form [4, 25, 37]. In the Laine family, the recessive variant was c.884C>A, p.(Ser295Ter); in the Pyott families, the variants included a homozygous missense variant, a homozygous nonsense variant, a nonsense variant with a single-nucleotide duplication in trans and a homozygous 14-base-pair deletion, with predicted mRNA decay and absence of WNT1 [42, 48]. Monoallelic forms cause autosomal dominant osteoporosis (OMIM 615221): the Finnish variant NM_005430.4:c.652T>G, NP_005421.1:p.(Cys218Gly), NC_000012.12:g.48981179T>G, was found in two large families with 25 carriers aged between 11 and 76 years [37, 42, 50].

The altered WNT1 proteins do not activate LRP5-mediated β-catenin signalling and induce fewer target genes and less mineralisation in vitro; osteoblasts increase Wnt1 expression with differentiation, and in the mouse Wnt1 is expressed in the bone marrow (B lineage and haematopoietic progenitors) and in a subset of osteocytes, suggesting communication between the haematopoietic and osteoblastic lineages [42, 47]. In the mouse, loss of Lrp5 did not abolish the anabolic effect of Wnt1, so this co-receptor does not appear to be indispensable for WNT1 function; in patients, deficient WNT1 mainly reduces cortical thickness through lower periosteal formation [5]. In a Chinese cohort of 16 patients with 16 variants, biallelic nonsense or frameshift variants were associated with earlier fractures and a more severe phenotype, and β-catenin was reduced in the bone of a compound heterozygous patient [63]. The combination of rare variants in several genes in the same patient, for example in WNT1, PLS3 and NOTCH2, has been described in a Brazilian cohort and makes interpretation more difficult [37, 64].

2.5 PLS3: X-linked osteoporosis

PLS3 is located at Xq23, with the MANE transcript NM_005032.7, and encodes plastin 3, a protein that bundles filamentous actin [43, 50]. Pathogenic variants are loss-of-function variants: nonsense, frameshift, splicing, missense, intragenic or whole-gene deletions and one partial duplication [37]. The proposed mechanism involves mechanosensing by osteocytes, whose processes are rich in actin, and mineralisation, because plastin 3 is present in the microvilli from which matrix vesicles bud [37]. Biopsies show low-turnover osteoporosis, sometimes with increased unmineralised osteoid, and high-resolution peripheral quantitative computed tomography (HR-pQCT) shows predominantly cortical deficits, more marked at the distal radius than in OI type I [37, 65]. The Pls3−/− mouse only reproduces the cortical phenotype [37]. In the Dutch cohort of 28 patients from 11 families, each family with a different variant, the type and location of the variant did not predict severity [6]. A rare variant, rs140121121, doubled fracture risk in heterozygous older women, which also makes PLS3 a factor in common multifactorial osteoporosis [43]. Missense variants in the actin-binding domains cause, through another mechanism, an allelic X-linked congenital diaphragmatic hernia [16].

2.6 SGMS2: a defect of sphingomyelin metabolism

SGMS2, at 4q25, encodes sphingomyelin synthase 2 (SMS2), a plasma membrane enzyme expressed mainly in cortical bone; the MANE transcript is NM_001375905.1, and the recurrent variant NM_001375905.1:c.148C>T, NP_001362834.1:p.(Arg50Ter), NC_000004.12:g.107895701C>T, has the same numbering in transcript NM_152621.6 used in the original publications [33, 50]. The nonsense variant p.(Arg50Ter) produces a catalytically inactive enzyme; the missense variants c.185T>G, p.(Ile62Ser), and c.191T>G, p.(Met64Arg), are retained in the endoplasmic reticulum and increase de novo sphingomyelin synthesis [33]. Osteoclast differentiation and function in vitro are normal, and biopsies show a mineralisation defect, reduced mineral content, heterogeneous mineralisation and severely disrupted lamellarity, with a woven-bone appearance [33, 37]. Transcriptomics of patient fibroblasts showed 215 upregulated and 58 downregulated genes, involving interleukin, gap junction and circadian clock pathways, and serum lipidomics showed 55 increased lipids, mainly triacylglycerols, mechanistic data not yet translated into clinical practice [66]. An Indian family with a frameshift duplication (c.1052_1074dup) presented fragility fractures with diffuse high bone mass, which broadens the spectrum [67]. In Saudubray's review, heterozygous pathogenic SGMS2 variants had been described in six families with early-onset osteoporosis, and the mechanism by which sphingomyelin synthase isoform 2 governs bone homeostasis and mineralisation remained unclear [68].

2.7 Inheritance, penetrance and expressivity

OPPG is autosomal recessive; carriers have reduced bone mass, but rarely evident eye disease: of the 10 parents with formal ophthalmological examination in the series of Ai, 2 had subtle retinal signs [27]. In FEVR, heterozygous LRP5 variants behave as dominant with incomplete penetrance and are associated with reduced BMD: 5 of 7 patients with LRP5-related FEVR had BMD more than 1 standard deviation below the mean, compared with 0 of 5 with FZD4 variants [27, 52]. A family with FZD4 and LRP5 variants in cis had a more severe ocular phenotype, suggesting complex inheritance [52]. In osteoporosis due to WNT1, LRP5 and SGMS2, inheritance is autosomal dominant with highly variable expressivity, including carriers without a phenotype [5, 8]. PLS3-related osteoporosis has complete expression in men and variable expression in heterozygous women, although a girl with a de novo variant had a Z-score of −6.6 at 6 years of age [6, 36]. The monogenic forms coexist with a polygenic component: children with a significant fracture history without a monogenic cause have an increased burden of common risk alleles [37].

3. Clinical Spectrum and Manifestations by System

3.1 Osteoporosis-pseudoglioma syndrome

Eye. Ocular involvement is congenital or infantile and dominates initial recognition. In the series of Ai and colleagues, all 30 probands with LRP5 variants had congenital or childhood eye disease, with a spectrum ranging from phthisis bulbi to persistent hyperplastic primary vitreous, congenital retinal folds and exudative retinopathy; most were bilaterally blind from birth, some were blind in one eye with impairment in the other or had moderate bilateral impairment, and all adults with variants were blind by 25 years of age, most already by 15 [27]. Retrolental masses (pseudoglioma) may be associated with vitreous hyperplasia, FEVR, corneal opacities and secondary glaucoma; ultrasound, CT or MRI show retrolental masses and intraocular calcification [3]. The underlying substrate is failure of involution of the primary vitreous vasculature, which normally begins at around the 13th week of gestation, with subsequent fibrosis and contraction [3]. In the Mennonites, the minimal light perception present at birth was lost in the first year in the families with W425X homozygosity and by 10 years of age in the family with compound heterozygosity; ocular pain was frequent and one eye was enucleated at 7 years for pain, with phthisis bulbi, corneal and iris neovascularisation, retinal disorganisation with fibrosis and osseous metaplasia of the pigment epithelium [38]. GeneReviews includes retinal detachment and microphthalmia among the ophthalmological problems of OPPG, and the chapter on NDP-related retinopathies describes it as pseudoglioma with FEVR, usually severe and exudative, with congenital blindness [25, 26].

Skeleton. Osteoporosis is generalised, with thin cortices of the tubular bones, collapsed and biconcave vertebral bodies, multiple fractures and secondary deformities due to malunion [3]. Skeletal disease was recognised by adolescence in 29 of 30 probands and in almost all affected siblings, but in fewer than half in the first two years of life; several asymptomatic infants had osteoporosis or vertebral compressions discovered on radiographs taken for pneumonia or constipation, and in some patients the bone disease was only recognised in the third decade [27]. In the Mennonites, fractures began in early childhood; one patient had been wheelchair-bound since the age of 7 [38]. In Turkey, ocular symptoms appeared in the first months, but the diagnosis of OPPG was only made after the age of 4, because of deformities and osteopenia [69]. In the natural history preceding modern treatment, bone disease progresses with demineralisation, fractures, growth arrest and deformities, and many patients become wheelchair-dependent in the second decade [3]. Short stature is frequent but not universal (2 of 9 Mennonite patients below the 5th centile) [3, 38]. An atypical subtrochanteric femoral fracture has been described in a 38-year-old man with OPPG, with relatively low turnover and trabecular osteoporosis on histology [61].

Nervous system and other. Muscular hypotonia, ligamentous laxity, mild intellectual disability and seizures are part of the classic descriptions [51, 70]. Cognitive impairment independent of visual impairment was reported in 8 of 30 probands with variants (4 of 12 homozygotes, 2 of 14 compound heterozygotes, 2 of 4 with a single identified allele), unrelated to the type of variant [27]. In the Mennonites, intelligence was globally normal, the sclerae white and joint mobility normal; one child had autism spectrum disorder and mild hypotonia [38]. The male:female ratio was 3:2 among probands with variants [27].

Biochemistry. Calcium, phosphate, alkaline phosphatase and turnover markers are normal for age; vitamin D insufficiency and deficiency were frequent in the Mennonite patients [38].

3.2 Osteoporosis due to monoallelic LRP5 variants

Children and adults with heterozygous LRP5 variants have low BMD and fractures, without severe eye disease or with mild vitreoretinopathy [5]. In the Stürznickel cohort, 37 of 50 carriers had fractures (vertebral in 18 and peripheral in 29), all had low bone formation, 31 had a Z-score ≤ −2.0, spine values were lower than hip values (−2.1 ± 1.3 versus −1.6 ± 0.8; p=0.003) and HR-pQCT showed compromised trabecular and cortical microarchitecture; the phenotype of index cases was more severe than that of relatives, with no difference between variant classes [28]. LRP5 variants have been described in women with pregnancy- and lactation-associated osteoporosis [37]. The case of two North American siblings with a heterozygous p.(Arg494Gln) variant, repeated appendicular fractures, gracile bones, coxa valga, dental agenesis and subclinical FEVR detected by angiography shows the overlap between the bone and ocular phenotypes [34]. A Birmingham series of five children, referred at a median age of 9 years for recurrent low-impact fractures, included siblings homozygous for NP_002326.2:p.(Gly420Ala) (published protein designation), one compound heterozygote and one heterozygote with a VUS inherited from the mother, and proposed considering LRP5 whenever recurrent fractures coexist with visual involvement [71].

3.3 WNT1-related osteoporosis and OI

Monoallelic form. In the Finnish families with p.(Cys218Gly), eight children and young people (median 14 years; 10–30) had no dysmorphic features or features of OI, had normal biochemistry and turnover markers, at least one fracture each (four with a history of pathological fractures), radiographs with an osteopenic appearance, loss of vertebral height and thin diaphyses, BMD below the median in all and a more marked deficit in the older individuals; histomorphometry at 14 and 16 years showed low-turnover osteoporosis [31]. The fibulae are particularly thin, and vertebral compressions are rare in childhood and young adulthood and almost universal after the age of 50 [37]. On spinal MRI of 18 carriers, 78% (7/9) of those over 50 had vertebral compression fractures, none of the younger ones had them, and the total was 39%, compared with 14% (2/14) of non-carriers with a single mildly compressed vertebra; carriers had a spinal deformity index of 4.0 ± 7.3 versus 0.0 ± 0.4, increased thoracic kyphosis in 33% and Schmorl nodes in 61% (versus 36%) already in adolescence, with enlarged intervertebral discs [35]. Other findings include raised intact and C-terminal FGF23, normal sclerostin and DKK1, a distinct microRNA profile, marrow fibrosis and one case of myelofibrosis in a young adult [37]. Some carriers have normal or osteopenic BMD, no fractures and good quality of life, and others have marked fragility, hypotonia, hypermobility, motor delay, gastrointestinal symptoms, strabismus or neurological symptoms [5]. In adults, the disease may first present with back pain, multiple vertebral fractures, rapid height loss and progressive kyphosis, as in a 27-year-old man and his mother, both without skeletal symptoms in childhood [32]. In four Chinese children with monoallelic variants, all had multiple vertebral fractures, the first fracture occurred between 2 and 12 years of age and the height-adjusted lumbar Z-score ranged from −8.06 to −3.50 [72]. A 28-year-old Brazilian woman presented with vertebral and rib fractures in the immediate postpartum period, with a lumbar Z-score of −4.7, and had a pathogenic WNT1 variant [73].

Biallelic form. OI type XV is moderately severe and progressive, clinically difficult to distinguish from dominant OI type III, with early fractures, deformities, kyphoscoliosis and severe short stature [5, 48]. Neurological manifestations have been described (severe cerebellar hypoplasia, epilepsy, pontine and midbrain tectal hypoplasia, severe global delay, cerebral atrophy), more frequent in the biallelic forms, and, in the Chinese cohort, brain anomalies, haematological disorders and a pituitary adenoma [5, 63].

3.4 X-linked osteoporosis (PLS3)

Hemizygous men have severe, early and progressive osteoporosis, dominated by multiple vertebral compressions, with frequent peripheral fractures [37]. In the Dutch cohort, the 18 men (median 47 years) had a lumbar Z-score of −2.8 ± 1.7 and a femoral neck Z-score of −1.7 (interquartile range −2.9 to 0.8) and a median of 12 fractures; of the 10 women (median 49.5 years), most had normal BMD, two had osteoporosis and one osteopenia, with a median of 2 fractures; mildly blue sclerae occurred in 33.3% of men and 30% of women, joint hypermobility in 44.4% and 70% and skin hyperlaxity in 50% and 20%, creating overlap with mild OI; within a 10-member family the heterogeneity was considerable [6]. On MRI, vertebral changes predominate in the upper thoracic spine (T3 to T6), are already present in childhood, in both sexes and more so in men; the vertebrae are biconcave and the endplates are preserved, with fewer Schmorl nodes than in controls [74]. In a screening of 95 Finnish children with bone fragility, only two of the 31 with severe primary osteoporosis of unknown cause had PLS3 variants: a boy with bilateral femoral fractures at 10 years, a Z-score of −4.1 at 18 and multiple vertebral compressions, and a girl with a de novo variant, long-bone and vertebral fractures and a Z-score of −6.6 at 6 years; none of the 64 children with multiple fractures but otherwise healthy had variants [36]. Heterozygous women may have peripheral and vertebral fractures, mainly in adulthood and after the menopause [37, 65].

3.5 Bone fragility with calvarial “doughnut” lesions (SGMS2)

Carriers of p.(Arg50Ter) have childhood-onset osteoporosis, with or without cranial sclerosis, and vertebral and peripheral fractures; carriers of the missense variants p.(Ile62Ser) and p.(Met64Arg) have neonatal fractures, severe short stature and spondylometaphyseal dysplasia [33]. “Doughnut” lesions are radiolucent areas surrounded by a sclerotic ring in the calvaria, palpable as nodules, with increased uptake on scintigraphy, and appear in late childhood or adolescence; there may be osteosclerotic maxillary and mandibular lesions, osteopenia with coarse trabeculae, mild bowing of the long bones, diaphyseal undertubulation, cortical thickening, caries, poorly developed teeth, dentinogenesis imperfecta, myopia and glaucoma; alkaline phosphatase may be moderately raised [49]. Femoral bowing may be visible on prenatal ultrasound, and fractures recur in childhood and decrease after puberty [49]. The neurological manifestations are distinctive: recurrent, transient palsies of the facial, trochlear and oculomotor nerves, isolated and spontaneously remitting, deafness, sensory neuropathy with progressive ataxia and global delay, without structural changes on brain imaging, suggesting a direct effect of neuronal sphingomyelin metabolism [5]. Inter- and intrafamilial variability is wide, ranging from absence of a bone phenotype to severe vertebral compressions; two probands with p.(Arg50Ter) were obese and had impaired muscle function, with low cortical vBMD and thin cortices on peripheral tomography [8, 29].

3.6 Idiopathic juvenile osteoporosis

IJO presents in the prepubertal period, between 8 and 12 years of age, earlier in girls, although more boys are affected, with pain mainly in the lower back and the joints of the lower limbs, walking difficulty, fragility, thoracolumbar kyphosis, a prominent sternum and a slight reduction in trunk height; there are metaphyseal fractures of the femora and tibiae and biconcave vertebral collapse; calcium, phosphate, vitamin D metabolites and PTH are normal, with a negative calcium balance in the acute phase [9]. After puberty, usually within 2 to 4 years, most recover without major clinical or radiological sequelae; a minority remain disabled in adulthood, with deformities and a final height equal to that at onset [9]. A patient in Spranger's atlas, initially diagnosed with OI because of knee pain at 7 years of age, improved from the age of 13, was well at 28 after surgical correction of the deformities and had healthy children [9]. Some of the so-called idiopathic osteoporoses have a monogenic cause not detected by limited testing [37].

3.7 Comparative overview

Sources: [5, 9, 27, 28, 35, 37, 49, 74].

4. Clinical Recognition and Warning Signs

4.1 Context: fractures in childhood

One third to one half of children sustain a fracture during childhood and 20% sustain two or more, mainly in the period of rapid growth that precedes and accompanies puberty; an isolated fracture is not enough to suspect bone disease [37, 70]. The clinical approach includes a three-generation pedigree, with questions about fractures, blue sclerae, dental problems, deafness, short stature, height loss and kyphosis; the age at each fracture, the bones involved and the energy of the trauma; the consistency of the history, because child abuse must be considered; healing, deformities, dislocations, hearing and teeth [70]. Examination assesses stature, deformities, sclerae, scoliosis, ligamentous laxity by the Beighton score (about 10% of the population scores above 5), skin and teeth [70].

4.2 When to suspect a monogenic osteoporosis

DXA is indicated, in children and in adults under 50 years of age, in the presence of two or more fragility fractures before the age of 10; more than two fragility fractures at any age; a fracture at an unusual site (spine, hip); more than two fractures with a family history of fractures or with extraskeletal signs of OI; or a fragility fracture in a patient with chronic disease or on glucocorticoids [37]. In the absence of a secondary cause, suspicion of a monogenic form increases with very early onset and with a family history of osteoporosis [5]. Some signs point to a particular gene. Congenital or infantile blindness, retrolental masses, persistent hyperplastic primary vitreous, retinal folds or exudative retinopathy in an infant or child with fractures or radiological osteoporosis suggest OPPG. Childhood visual loss without vitreoretinal disease makes LRP5 unlikely [27]. FEVR, even mild or only angiographic, in a child with fractures, or a family with FEVR and fractures, suggests heterozygous LRP5 [34, 52]. Vertebral and long-bone fractures in a boy, with a mother or sister with adult-onset osteoporosis, suggest PLS3 [6, 36]. Long-bone fractures in childhood with thin fibulae and, in adult relatives, multiple vertebral compressions, kyphosis and height loss after the age of 50, suggest WNT1 [31, 35]. Palpable bony nodules on the skull, ring-shaped sclerotic lesions of the calvaria and episodes of transient peripheral facial palsy suggest SGMS2 [37, 49]. Recurrent fractures with visual involvement of any degree justify testing of LRP5 [71].

4.3 Approach to retinal dysplasia or pseudoglioma

Congenital bilateral retinal dysplasia requires exclusion of retinoblastoma, severe FEVR and, in the preterm infant, retinopathy of prematurity; the history should include fractures (OPPG), and the work-up includes a full ophthalmological examination, brain MRI, creatine kinase (raised in dystroglycanopathies such as Walker-Warburg syndrome) and genomic analysis [51]. Retinoblastoma must be excluded as the cause of retrolental masses [3]. In infants and children with ocular findings of OPPG, Ai and colleagues recommended radiographs and quantitative BMD measurement, because vertebral compressions may be clinically silent and early diagnosis allows the bone disease to be treated [27].

4.4 Warning signs and emergencies

Back pain in a child or adolescent: may correspond to vertebral compression fractures, which are often asymptomatic or attributed to another cause [9, 75]. Low-trauma femoral fracture, or prodromal pain in weight-bearing bones during prolonged bisphosphonate treatment: consider atypical femoral fracture [61, 76]. Ocular pain in a patient with OPPG: secondary glaucoma or painful phthisis, which may require enucleation [3, 38]. Vision still present in a child with LRP5-related FEVR: a retinal hole or avascular areas require urgent assessment by a retinal specialist and possible laser treatment to prevent detachment [34]. Fractures without a plausible explanation: consider child abuse in parallel with investigation for bone disease, without one excluding the other [25, 70].

None of the sources consulted describes newborn screening programmes for these entities; early recognition depends on ophthalmological assessment of the blind infant and on the investigation of fractures.

5. Diagnosis: Criteria, Tests and Algorithms

5.1 Criteria for osteoporosis in children and young adults

The 2013 paediatric official positions of the International Society for Clinical Densitometry (ISCD) revised those of 2007 and are organised into four positions [77]. The first sets the goal of bone health assessment: to identify children and adolescents who may benefit from an intervention to reduce an elevated risk of clinically significant fracture [77]. The second establishes that one or more vertebral compression fractures, defined as a loss of height greater than 20% at any point of the vertebra, indicate osteoporosis in the absence of local disease or high-energy trauma, and that BMD measurement, in that case, only completes the assessment (grade B, fair evidence) [77]. The third requires, in the absence of a vertebral fracture, the combination of a clinically significant fracture history (two or more long-bone fractures by 10 years of age, or three or more at any age up to 19 years) with a BMD or bone mineral content (BMC) Z-score ≤ −2.0, and adds that a Z-score > −2.0 does not exclude skeletal fragility or increased fracture risk, as seen in mild OI (grade C, poor evidence) [77]. The fourth maintains that the diagnosis of osteoporosis in children and adolescents is not made on densitometric criteria alone (grade C) [77]. Compared with 2007, one lower-limb long-bone fracture and two upper-limb long-bone fractures at any age ceased to be sufficient criteria, because the working group considered these thresholds too permissive: applied as a screening criterion, they would lead up to 34% of children to undergo DXA [77]. The new numerical thresholds derive from those of Mäyränpää and colleagues (two fractures before 10 years of age, three by 16 years or one vertebral fracture); the working group acknowledged that no data show these numbers or ages to predict risk better than others, and the final version is an expert consensus [77]. Fractures due to high-energy trauma, such as road traffic accidents and falls from more than 3 m, are excluded; the distinction between mild and moderate trauma was left to clinical judgement, for lack of data, and the Landin criteria were rejected because they classify skiing and skateboarding falls as mild [77]. In the Mäyränpää series, children with a vertebral fracture had a mean lumbar Z-score of −0.8 ± 0.8, and few had an adjusted Z-score < −2.0, which supports the sufficiency of a vertebral fracture without a densitometric threshold [77]. The working group restricted the Z-score to the spine and total body, because hip data are not reliable in young children, found no published prediction equation adjusting bone mass for height, weight and bone area, and found no data justifying different thresholds by ethnicity or in obesity [77]. The 2018 consensus of the Australasian Paediatric Endocrine Group (APEG) adopts the same definition, with a BMD Z-score < −2 using norms adjusted for age, sex and height and recurrent long-bone fractures, or an isolated vertebral compression fracture, and points out that recurrent long-bone fractures may occur with normal BMD [78]. In the formulation followed by the reviews of Costantini and of Mäkitie and Zillikens, the diagnosis in children requires: (a) one or more vertebral compression fractures without relevant trauma, a sufficient criterion even with normal BMD; or (b) a Z-score ≤ −2.0 at the lumbar spine or total body, adjusted for height and pubertal stage, with a significant fracture history, defined as two or more long-bone fractures before 10 years of age or three or more before adulthood, counting only low- to moderate-energy fractures and excluding fractures of the fingers, nose and similar [5, 37]. This formulation differs from the ISCD in the age limit for the three fractures (adulthood in the reviews, 19 years in the ISCD) and in the explicit exclusion of finger and nose fractures, which the ISCD does not mention [5, 37, 77]. Once adult height has been reached and up to 50 years of age, osteoporosis is defined by a Z-score < −2.0 (or T-score < −2.5) at the spine or femoral neck associated with fragility fractures or chronic disease [5, 37]. In children, low BMD without fractures or underlying disease is not equivalent to osteoporosis and is associated with a low short-term fracture risk [37].

There are no consensus diagnostic criteria for OPPG; the minimum criteria proposed by Frontali and colleagues (vitreoretinal dysplasia or phthisis bulbi with radiological osteoporosis) remain useful as a clinical definition [44]. IJO is only diagnosed after secondary causes and known monogenic causes have been excluded [9, 37].

5.2 Exclusion of secondary causes

Mäkitie and Zillikens propose a two-level laboratory work-up [37]:

In children, baseline biochemistry should include calcium, phosphate, magnesium, alkaline phosphatase, PTH, 25-hydroxyvitamin D and urinary calcium/creatinine ratio, and normal calcium, vitamin D and bone profile are a prerequisite for treatment [75]. Low alkaline phosphatase points to hypophosphatasia, and urinary amino acid testing to homocystinuria [70]. In the monogenic forms of the class, conventional biochemistry is usually normal, which does not exclude the diagnosis [31, 38].

5.3 Imaging and densitometry

DXA is the preferred method; in children, total body less head and the lumbar spine in the posteroanterior projection are measured, and the 2019 ISCD positions added the proximal femur, lateral distal femur, radius and DXA vertebral fracture assessment (VFA) when reference data exist [75, 79]. DXA measures areal density and underestimates BMD in small children or those with growth delay, so it is corrected for stature, with bone mineral apparent density (BMAD) on Lunar devices or the height-adjusted Z-score on Hologic devices [75]. In the Mennonites, the authors pointed out that paediatric Z-scores do not exist for the spine before 3 years of age or for the hip before 5 [38]. Lateral spine radiography diagnoses vertebral fractures, with limitations in the thoracic spine; the low-dose EOS system is an alternative [75]. The APEG consensus considers lateral thoracolumbar imaging (radiography or DXA vertebral morphometry) an integral part of the assessment of fragility, to be performed at baseline and when clinically indicated, and DXA the only available surrogate measure of bone accrual, to be performed before the first bisphosphonate dose and annually, with the limitations of reference data below 3 years of age and of contractures or orthopaedic hardware; peripheral quantitative computed tomography (pQCT) complements DXA in experienced centres, and turnover markers (osteocalcin, procollagen type 1 N-terminal propeptide, collagen cross-links, deoxypyridinoline) remain mainly research tools [78]. A radiological skeletal survey documents old fractures and is important for distinguishing genetic disorders from child abuse; skull radiography looks for wormian bones [70]. HR-pQCT distinguishes cortical (for example, in PLS3-related osteoporosis) and trabecular patterns, but is mainly a research tool [28, 65]. In OPPG, ocular ultrasound, CT or MRI show the retrolental masses and intraocular calcification [3]. In FEVR, fluorescein angiography reveals peripheral avascular areas that may be the only manifestation in mildly affected relatives [34, 52].

Transiliac bone biopsy, with prior tetracycline labelling, is the only site with paediatric reference values and provides information on the type of osteoporosis (low or high turnover), the presence of osteomalacia and rare causes such as mastocytosis or myeloma; it is invasive, requires expertise and is reserved for unresolved cases or research [5, 75]. In the forms of this class, biopsies have shown low turnover in WNT1 and PLS3, reduced cortical width and trabecular volume in LRP5 and WNT1, increased mineralisation lag time in PLS3 and woven bone in SGMS2 [5].

5.4 Molecular diagnosis

The recommended strategy is sequencing first, supported by robust clinical and biochemical characterisation [5]. Gene panels for OI and monogenic osteoporosis should include COL1A1 and COL1A2, because OI may not have the typical appearance, and genes such as LRP5, LRP6, WNT1, PLS3, SGMS2, DKK1 and those of the recessive forms of OI [37, 62]. Targeted panels (50–500 kb) give good coverage; the exome (about 45 Mb) allows reanalysis and discovery of new genes, but has intronic gaps; the genome also detects non-coding and structural variants [5]. Trio sequencing (proband and parents) increases the ability to exclude variants that do not follow the inheritance pattern [5]. When FEVR is suspected, the non-syndromic panel includes CTNNB1, FZD4, LRP5, TSPAN12 and ZNF408, and a negative panel does not exclude the diagnosis, because only about half of families with FEVR have an identified variant [26]. In SGMS2-related disease, the recurrent variant p.(Arg50Ter) can be tested in a targeted way [29].

Limits of testing. Detection by exon sequencing was about 70% of the expected alleles in the 2005 OPPG series [27]. Conventional sequencing misses deletions of one or more exons (for example, the homozygous deletion of exon 4, the 7.2 kb deletion and the deletion of exons 14 to 16 due to Alu recombination), splicing variants whose effect can only be demonstrated in RNA and deep intronic variants [3, 58, 59]. As bone genes are poorly expressed in leucocytes, RNA has to come from skin biopsy fibroblasts [5]. Copy-number and mosaicism analysis should be requested or confirmed in the report. A negative result does not exclude a monogenic form, and in a Brazilian cohort of 37 patients with idiopathic, severe or familial osteoporosis, a 128-gene panel found only three pathogenic or likely pathogenic variants (in COL1A2, WNT1 and IDUA) [64]. In a Finnish cohort of 44 patients with idiopathic early-onset osteoporosis there were no pathogenic SGMS2 variants [80]. The proportion of detectable variants in a selected series is not the clinical sensitivity of the test.

Interpretation. Sequencing of an exome reveals about 90,000 single-nucleotide variants, filtered by frequency, predicted pathogenicity, inheritance and phenotype (HPO ontology) [5]. VUS are frequent, especially in LRP5: 14 of the 50 relevant variants in the Stürznickel cohort were VUS and 19 were BMD-associated polymorphisms [28]. In LRP5, interpretation includes the direction of effect, because the same gene causes low and high bone mass disorders; WNT and norrin signalling reporter assays distinguished OPPG and high bone mass mutants, but are not available in routine practice [5, 27]. Family segregation and the phenotype of relatives are essential, with the caveat that relatives may have low BMD for non-genetic reasons [37]. The report should state the gene, the transcript with version (for example, NM_002335.4 for LRP5, NM_005430.4 for WNT1, NM_005032.7 for PLS3 and NM_001375905.1 for SGMS2), the HGVS c. and p. nomenclature, zygosity, the ACMG/AMP classification and technical limitations [50]. In articles published before 2010, LRP5 variants appear in abbreviated protein nomenclature (W425X, T409A, R1036Q, N531I), counted from the initiating methionine [27].

5.5 Algorithm

1. Fragility fractures, vertebral fracture or low BMD with fractures → confirm the criteria for osteoporosis (5.1) and exclude child abuse when appropriate [5, 70].
2. Exclude secondary causes with the laboratory and clinical work-up (5.2) and correct calcium and vitamin D deficiencies [37].
3. Look for pointers: eye (OPPG, FEVR), skull and cranial nerves (SGMS2), sex and X-linked family history (PLS3), late vertebral compressions in the family (WNT1), signs of OI (blue sclerae, dentinogenesis imperfecta, deafness, wormian bones) [5, 37].
4. Ophthalmological examination with fundoscopy and, if in doubt, fluorescein angiography in all children with primary osteoporosis and in relatives of patients with LRP5 variants [27, 34].
5. Bone fragility panel with CNV analysis, or exome/genome with filtering, preferably as a trio; segregation in relatives [5].
6. Negative result with a suggestive phenotype: periodic reanalysis, genome, RNA studies or bone biopsy at a reference centre; only then a diagnosis of idiopathic osteoporosis [5, 37].

5.6 Prenatal and preimplantation diagnosis

When the familial variants are known, carrier testing and prenatal or preimplantation diagnosis are technically possible [51]. In severe forms, ultrasound may show bowing of the long bones (described in SGMS2-related disease), but the ocular changes of OPPG have no described prenatal expression in the sources consulted [49].

6. Differential Diagnosis

6.1 Comparative table

6.2 Main differentials in prose

OI versus OPPG and non-collagen monogenic forms. OI is distinguished from OPPG by the absence of visual impairment and by the presence of wormian bones and blue sclerae [3]. Without ocular involvement, OPPG may mimic a progressively deforming or moderate OI [4]. The distinction between mild OI and PLS3-related osteoporosis is difficult, because one third of patients with PLS3 variants have mildly blue sclerae and many have hypermobility; DXA does not separate them, but HR-pQCT showed a more marked cortical deficit in PLS3-related osteoporosis [6, 65]. The presence of dentinogenesis imperfecta, deafness or wormian bones favours OI [70]. IJO begins in mid to late childhood, affects the axial skeleton more than the extremities and is not associated with wormian bones or ocular or dental defects, which separates it from OI type I [82].

OPPG versus Norrie disease and FEVR. Low bone density and fractures distinguish OPPG from Norrie disease, which shares the ocular picture [26]. Norrie disease affects boys, has bilateral retrolental masses without intraocular calcification, progressive sensorineural deafness in at least 40% and learning difficulties in about half, and female carriers rarely have a phenotype [51]. In the series of Ai, probands without LRP5 variants had a boy:girl ratio of 6:1, which raises the hypothesis of X-linked causes, although NDP was normal in all [27]. LRP5-related FEVR may be associated with low BMD and fractures, unlike FZD4-related FEVR, and some authors regard LRP5-related FEVR and OPPG as parts of a single spectrum [27, 34].

OPPG versus spondyloocular dysplasia. Both combine bone fragility and eye disease; XYLT2-related dysplasia has cataract, lens malformation, retinal detachment, deafness, heart disease and facial dysmorphism, ophthalmological findings distinct from those of OPPG [81].

SGMS2-related disease versus gnathodiaphyseal dysplasia. Gnathodiaphyseal dysplasia has more extensive mandibular lesions, histologically multiple cemento-ossifying fibromas, diaphyseal sclerosis and fragility, but not calvarial “doughnut” lesions [49].

IJO versus monogenic forms. IJO has prepubertal onset and spontaneous remission; OI differs by its earlier onset, more diaphyseal fractures, narrow tubular bones, wormian bones and blue sclerae; adult idiopathic osteoporosis affects older patients, with rare long-bone fractures [9]. LRP5 variants have been found in children with “idiopathic juvenile osteoporosis”, so exclusion of the monogenic forms requires genetic testing [30, 37, 38].

The opposite pole of LRP5. Gain-of-function variants cause Worth-type endosteal hyperostosis, mild autosomal dominant osteopetrosis and familial high bone mass with torus palatinus, without described eye disease [38, 53]. Classification of an LRP5 variant therefore requires consistency between the predicted direction of effect and the phenotype.

7. Treatments and Risk Management Strategies

7.1 Principles and state of the evidence

There is no evidence-based guideline for early-onset osteoporosis or for any of the monogenic forms of this class; the closest document is the 2018 APEG consensus on bisphosphonates in children and adolescents, which addresses primary osteoporosis on the basis of OI [78], and no drug has an approved indication for OPPG, for LRP5-, WNT1-, PLS3- or SGMS2-related osteoporosis, or for paediatric osteoporosis [5, 37, 75]. The available studies are small series, case reports and retrospective cohorts, almost always with BMD as the outcome rather than fractures [37]. Data from collagen-related OI cannot be extrapolated directly, because bone tissue and turnover differ: several forms of this class (WNT1, PLS3, LRP5) are low-turnover, a situation in which an antiresorptive agent may bring limited benefit [5]. Mäkitie and Zillikens recommend a personalised approach in which absence of evidence is not taken as evidence of absence of effect, led by a multidisciplinary team experienced in paediatric and young adult bone disease [37]. The doses given below come from published series in the disease itself when available, from the APEG consensus (reference) and paediatric reviews for bisphosphonates, and from official labels for comparison; all correspond to off-label use.

7.2 General measures

Calcium and vitamin D are first-line when diet or biochemistry shows insufficiency; there are no fracture data, but small studies show an increase in BMD [5, 37]. In children, 25-hydroxyvitamin D should be kept above 50 nmol/l, because insufficiency blunts the response to bisphosphonates and increases the risk of hypocalcaemia after infusion [75]. In the Mennonite OPPG series, all patients received calcium 1000 mg/day and colecalciferol 400–800 IU/day, titrated to 25-hydroxyvitamin D above 30 ng/ml, and vitamin D deficiency was frequent [38]. The GeneReviews text on OI advises against very high doses of calcium and vitamin D [25]. Lifestyle measures include an adequate diet, regular weight-bearing exercise, avoidance of alcohol and tobacco and a healthy weight; in children with recurrent fractures, low calcium intake, vitamin D deficiency and insufficient physical activity are major factors, to be corrected before any drug [5, 25, 37]. In the Dutch PLS3 cohort, the authors hypothesised a protective role for weight-bearing exercise [6]. Contact and high-impact sports, such as horse riding and trampolining, which may cause or worsen vertebral fractures, should be avoided [75]; the two siblings with LRP5 variants in the Littman case gave up contact sports and had no new fractures in 2 years [34].

Rehabilitation includes physiotherapy and occupational therapy for mobility and independence, early mobilisation after each fracture, assistive devices and pain assessment [75]. Fractures heal well in IJO and in SGMS2-related disease and should be immobilised for as short a time as possible [9, 49]. In OPPG, fractures require immobilisation and, in cases of displaced femoral fracture, internal fixation and intramedullary rodding [3]. Blind children with OPPG need adapted stimulation and education and ophthalmological follow-up for pain and glaucoma; a painful eye may require enucleation [3, 38]. In LRP5-related FEVR with preserved vision, prophylactic laser photocoagulation of avascular areas and retinal holes preserved vision in the published case [34].

7.3 Indication for pharmacological treatment

In children, the indication for treatment arises from the presence of fractures and persistent risk, not from the BMD value, which serves mainly for monitoring; the decision weighs severity (number of fractures and degree of vertebral compression), the remaining growth potential, because vertebral reshaping depends on growth, and the persistence of risk factors [75]. Children have the capacity to regain BMD and to reshape fractured vertebrae, sometimes even without treatment, and reshaping is better the earlier the bisphosphonate is started [75]. In OPPG, Streeten and colleagues recommended starting bisphosphonates in the first years of life, and Ai and colleagues insisted on early diagnosis of the bone disease to allow treatment [27, 38]. In young adults, when secondary causes have been excluded and fragility fractures continue, antiresorptive or anabolic agents are considered, taking pregnancy plans into account [5, 37]. In the Littman case, with traumatic fractures, no vertebral fractures and puberty beginning, the team deferred drug treatment until there was progression, which illustrates that not every low BMD with an LRP5 variant requires immediate treatment [34].

The APEG consensus grades its recommendations using the GRADE system (recommendation 1, strong, or 2, weak; quality of evidence from very low to high), and its algorithms for primary osteoporosis, built on OI, apply by extrapolation to the monogenic forms of this class [78]. In primary osteoporosis, a bisphosphonate is considered when there is significant evidence of fragility, that is, vertebral compression fractures or two low-trauma long-bone fractures; without these, bone health is only optimised (dietary calcium, sufficient vitamin D, weight-bearing exercise) [78]. In OI, IV bisphosphonates should be considered in severe forms (type III), in children with vertebral compression fractures and in those with two or more long-bone fractures per year, and oral agents only in mild to moderate forms without vertebral fractures (strong recommendation, moderate evidence) [78]. In IJO, it is recommended to consider bisphosphonates in severe forms, defined by two or more long-bone fragility fractures or by vertebral fractures (strong recommendation, low evidence); the basis is a single randomised trial of pamidronate with 5 treated patients, which reported fewer fractures and less pain [78]. Prophylactic treatment of a low Z-score without fractures is not recommended, and the drug is only started after vitamin D, calcium intake, mobility and pubertal delay or arrest have been corrected (strong recommendation, low evidence, formulated for secondary osteoporosis) [78]. The Cochrane review cited by the consensus (14 trials, 819 participants with OI, 2003–2013) showed an increase in BMD in all trials, with incomplete data and no power for growth, pain, fractures and function [78]. Treatment should be supervised by a paediatrician experienced in genetic bone disease, at a specialist centre with a multidisciplinary team [78].

7.4 Bisphosphonates

Mechanism and experience. Nitrogen-containing bisphosphonates (pamidronate, zoledronic acid, alendronate, risedronate, neridronate) inhibit osteoclasts and reduce turnover; they are the main pharmacological treatment for paediatric bone fragility and the standard of care in the United Kingdom, but are off-label worldwide, except for neridronate in OI in Italy [25, 75]. The 2018 consensus of the Australasian Paediatric Endocrine Group (APEG) presents GRADE-graded recommendations for primary bone fragility disorders and for secondary osteoporosis [78].

Reference regimen: 2018 APEG consensus (analogue, off-label). The consensus acknowledges that the ideal drug, dose and frequency have not been determined and that regimens vary with the resources and experience of each centre; the OI data refer mainly to pamidronate, with growing experience with zoledronic acid [78]. Pamidronate is often used below 2 years of age, with a switch to zoledronic acid in older children with moderate to severe OI [78]. The algorithm for primary osteoporosis is summarised as follows [78]:

In severe OI, the consensus advises against stopping when BMD improves: a prolonged low dose is maintained to preserve bone strength during growth, because bone formed after discontinuation does not benefit from previous treatment; in less severe forms it may be possible to stop during childhood and reintroduce when BMD begins to fall, a matter on which the consensus itself acknowledges a lack of evidence [78]. In IJO, given the lack of data on duration, it recommends reducing the dose after 2 years if height-adjusted DXA values are normalising, following the same algorithm [78]. In secondary osteoporosis, the algorithm provides for 12 months at the same doses, reassessment and, at 2 years, discontinuation or maintenance if fractures, pain or risk factors persist [78]. Oral bisphosphonates are only supported in mild to moderate OI and require the child to swallow the tablet whole with a glass of water and not to have gastro-oesophageal reflux, because of the risk of erosive oesophagitis (strong recommendation, very low evidence) [78].

Alternative regimen: approach of Sakka and Cheung (analogue, off-label). The authors stress that there is no consensus on dose, frequency and duration [75]. The review by Sakka and Cheung gives the following doses, used mainly in OI and applied by extrapolation to the monogenic forms of this class [75]:

In the authors' approach to primary osteoporosis, the full dose of zoledronic acid depends on age: under 2 years, 0.025 mg/kg (maximum 2 mg) every 3 months; 2 to 5 years, 0.035 mg/kg (maximum 2 mg below 3 years and 4 mg above) every 4 months; over 5 years, 0.05 mg/kg (maximum 4 mg) every 6 months. The first infusion is given as an inpatient, at half the dose and with prophylactic paracetamol; before each infusion, 25-hydroxyvitamin D ≥50 nmol/l and normal calcium and renal function are confirmed; the full six-monthly dose is maintained until BMAD exceeds 0 SD, then half the six-monthly dose is given and, above +2 SD, a quarter of the dose annually; treatment ends at skeletal maturity [75].

Discrepancies between dosing sources. The annual dose of zoledronic acid coincides above 5 years of age (0.05 mg/kg every 6 months equals 0.1 mg/kg/year), but the maximum differs: 4 mg per infusion in Sakka and Cheung and 5 mg per year in APEG [75, 78]. Below 2 years of age, Sakka and Cheung use three-monthly zoledronic acid, whereas APEG describes pamidronate as the usual drug at that age [75, 78]. The reduced first dose is half the dose in Sakka and Cheung and corresponds to a quarter or half of the six-monthly dose in APEG (0.0125 or 0.025 mg/kg) [75, 78]. Dose reduction is guided by the height-adjusted BMD Z-score in APEG (half between −2.0 and 0; maintenance at 0.025 mg/kg/year above 0) and by BMAD in Sakka and Cheung (half the six-monthly dose above 0) [75, 78]. Above +2 SD, APEG stops the bisphosphonate, whereas Sakka and Cheung switch to a quarter of the dose, once a year, until skeletal maturity [75, 78]. The regimens published in OPPG use pamidronate at doses and intervals different from both [38, 83]. None of these regimens has been compared with another in a trial, and all fall outside the conditions of the labels.

Doses published in OPPG. In the Mennonite series, oral risedronate (35 mg/week or 25 mg twice weekly) produced a minimal response, and patients were switched to IV pamidronate 1 mg/kg over 3 hours, on 3 successive weeks, repeated every 3 months, following the OI regimen; the youngest received oral alendronate 1 mg/kg/day [38]. In the Brazilian siblings from São Paulo, IV pamidronate was given at 1 mg/kg/day on 2 consecutive days, every 3 months in the first year and every 4 months thereafter, with calcium, phosphorus, alkaline phosphatase, PTH, transaminases, creatinine and full blood count before each cycle [83]. These regimens differ from those of the reviews and show the variation between centres.

Results in OPPG. Zacharin and Cundy treated 3 children aged 9 to 11 years with multiple vertebral collapses with intermittent IV bisphosphonates for 2 years (pamidronate in 2, clodronate in 1): bone pain and mobility improved early, radiographs showed dense new bone at the endplates and vertebral reshaping, and the mean lumbar Z-score went from −4.5 to −2.8 [84]. In the Mennonites, the four treated patients (1.5 to 6.5 years of treatment) improved: Z-score from −6.4 to −2.4, from −4.7 to −0.4, from −2.9 to −0.5 and from −4.2 to −0.3; the most robust responses occurred in the youngest, with alendronate, and a child started at 2 years and 8 months after a femoral fracture did not fracture again in 17 months [38]. In the Brazilian siblings, pamidronate was safe for up to 3 years, with increased BMD and fewer fractures [83]. In Turkey, 3 children treated for 3.5 to 7 years improved their lumbar Z-score and bone pain [69]. In Greece, 4 treated patients had less pain and better BMD, but 3 sustained a fracture during treatment [60]. In a Birmingham series, zoledronic acid (median 2 years) increased the lumbar BMAD Z-score by +1.0, +0.7 and +2.7 in three children, who did not fracture again, and by +0.3 in a compound heterozygous patient, with vertebral reshaping [71]. In adults, the effect tends to wane over time: a Mennonite patient had a femoral fracture on pamidronate and a plateau in Z-score, and a young Brazilian man had minimal effect in the last 3 of 6 years of pamidronate [38, 85].

Results in the other forms. In the Chinese WNT1 cohort, bisphosphonates and denosumab significantly increased spine and hip BMD and reshaped compressed vertebrae; in four children with monoallelic WNT1 variants, zoledronic acid increased lumbar BMD by 0.061 to 0.251 g/cm², reduced pain and there were no new fractures during short follow-up [63, 72]. In the Finnish adults with WNT1 variants, several developed extensive vertebral compressions despite years of conventional treatment, and in nine medication did not improve the vertebral pathology [5, 37]. In PLS3-related osteoporosis, adult men gained on average 16.6% in lumbar BMD (p=0.03) after a median of 6 years of treatment, mainly with oral bisphosphonates, and two Colombian brothers improved with zoledronic acid, calcium and vitamin D [6, 86]; Costantini's review adds that bisphosphonates appear to improve vertebral shape and prevent new fractures in this form [5]. In SGMS2-related disease, bisphosphonates in three children prevented new long-bone fractures and improved back pain, and prolonged treatment was associated with reshaping of compressed vertebrae [5, 8]. In IJO, bisphosphonates halt progression and accelerate spontaneous recovery, according to Spranger's atlas, which cites studies with pamidronate and clodronate [9]. In spondyloocular dysplasia, bisphosphonates did not reduce the fracture rate [81].

Adverse effects and management. An acute-phase reaction (fever, headache, myalgia, arthralgia, malaise, rarely vomiting for up to 72 hours) occurs with 18–27% of first IV infusions; paracetamol before and for 24–48 hours afterwards reduces the incidence by about half, and it decreases with subsequent infusions; transient hypocalcaemia and hypophosphataemia and a rise in CRP may occur, and supplemental calcium in the following days prevents hypocalcaemia when there has been a previous episode [75]. The APEG consensus describes a self-limiting acute-phase reaction in up to 80% of patients in the 24–48 hours after the first infusion, a frequency higher than that in the review by Sakka and Cheung; it recommends 25-hydroxyvitamin D > 50 nmol/l before the first infusion and adequate calcium intake afterwards, paracetamol and an antiemetic for symptoms and, as an option, calcitriol for 3 days after the first dose; severe symptomatic hypocalcaemia is rare, and hypophosphataemia does not require routine supplementation (strong recommendation, low evidence) [75, 78]. Before the first dose, APEG requires a full blood count, urea and electrolytes, liver function tests, 25-hydroxyvitamin D, PTH, calcium, magnesium and phosphate and a dental assessment, in addition to DXA and lateral spine imaging; during treatment, annual DXA when feasible, 25-hydroxyvitamin D and general biochemistry [78]. For a red or painful eye, the consensus requires ophthalmological examination to exclude iritis, described in isolated reports (strong recommendation, very low evidence), which in LRP5-related forms ties in with the disease's own ophthalmological follow-up [78]. Oral bisphosphonates mainly cause gastrointestinal intolerance [25]. Osteonecrosis of the jaw is rare in adults and has not been described in children treated with bisphosphonates [25, 75]; even so, APEG recommends dental assessment before the first dose, with invasive dental procedures completed beforehand, and dental review every 6–12 months during treatment (strong recommendation, low evidence) [78]. Sclerotic metaphyseal lines (“zebra lines”) record the treatment cycles, and the image of the OPPG patient in Spranger's atlas shows the resulting dense tibial metaphysis [3, 75]. Pamidronate has been associated with delayed healing of osteotomies, but not of fractures, and an infusion-free interval of a few months after an osteotomy is recommended [75]. APEG extends the precaution to fractures: in OI, whenever possible, it withholds the bisphosphonate until there is callus at the fracture or osteotomy site (strong recommendation, very low evidence), based on a case-control study of delayed osteotomy healing on pamidronate and despite another study, with a different surgical technique and zoledronic acid, having shown no delay [78]. Atypical femoral fractures have been described in OPPG, in a 38-year-old man with low turnover, and in an 18-year-old with a PLS3 variant after 7 years of IV and 2 years of oral bisphosphonates, with prodromal pain and a previous transverse tibial fracture that may belong to the same spectrum; these cases weigh on the decision about treatment duration in children with congenital bone disease [61, 76]. In OI, femoral fracture of atypical appearance relates more to disease severity than to bisphosphonate dose [75]. APEG considers that so-called atypical femoral fractures in treated young people may not be drug-related and are not, in themselves, an indication to stop (strong recommendation, low evidence), a position that contrasts with the weight given to the OPPG and PLS3 cases in the decision about treatment duration [61, 76, 78].

Labels (comparison). Pamidronate disodium has US indications for hypercalcaemia of malignancy (60–90 mg as a single IV dose in moderate hypercalcaemia), Paget disease and osteolytic metastases, with no osteoporosis or paediatric indication; the original US authorisation (Aredia®) dates from 31 October 1991 [87]. Oncology zoledronic acid (4 mg) is not indicated in children; in the 1-year paediatric trial in 152 children with severe OI (1–17 years; 74 treated; mean baseline lumbar Z-score −2.7), BMD increased, but individual changes did not correlate with fracture risk or pain; the 5 mg formulation for adult osteoporosis (Reclast® in the USA) was authorised on 16 April 2007 [88]. Alendronate (70 mg/week in adults) is not indicated in paediatrics; in the 2-year trial in 139 children with severe OI (5 mg/day below 40 kg, 10 mg/day from 40 kg; baseline Z-score −4.5), the lumbar Z-score rose by 1.3 versus 0.1 with placebo, with no reduction in fracture risk [89]. The paediatric regimens described (per mg/kg) are therefore outside the conditions of all the labels.

7.5 Teriparatide

Mechanism and rationale. Teriparatide is the recombinant 1-34 fragment of parathyroid hormone, with an osteoanabolic effect [25]. PTH stimulates the WNT pathway by reducing DKK1 and sclerostin, and in the Lrp5−/− mouse it increased bone formation as in the wild-type animal, mainly in cortical bone, which provided the rationale for its use in OPPG [38].

Dose (label). In adults, the approved dose is 20 µg subcutaneously once daily into the thigh or abdomen, with calcium and vitamin D if intake is insufficient [90]. In the USA, use for more than 2 years over a lifetime is only considered if fracture risk remains high; in the EU, treatment lasts up to 2 years and only one course is given in a lifetime [90, 91]. The indications are high-risk postmenopausal osteoporosis, high-risk primary or hypogonadal osteoporosis in men and glucocorticoid-induced osteoporosis [90, 91]. Paediatric safety and efficacy have not been established, children have an increased baseline risk of osteosarcoma because they have open epiphyses, and the European authorisation (Forsteo®) prohibits use in children and in young adults with immature bones [90, 91]; for this reason, teriparatide is not used in children with osteoporosis [75].

Results in the class. In a 19-year-old Brazilian man with OPPG, teriparatide 20 µg/day for 2 years, after 6 years of pamidronate, increased BMD by 9.7% at the lumbar spine and 10.2% at the hip, without adverse effects; CTX rose first (peak at month 3) and P1NP later (peak at month 9), both returning to baseline at 24 months, a pattern the authors attributed to low turnover [85]. In the most severely affected Mennonite patient, already an adolescent and with fractures on pamidronate, after 6 months without a bisphosphonate, teriparatide (20 µg on alternate days for 4 months and then daily for 11 months; the article gives the dose in mg, which is a typographical error) did not change the Z-score (−2.5 to −2.6), possibly because of prior suppression of turnover by bisphosphonates [38]. This use in an adolescent departs from the label precautions regarding immature skeletons and should not be taken as practice [90, 91]. In an uncontrolled prospective Finnish trial, six adults (median 54 years; three with WNT1 NM_005430.4:c.652T>G p.(Cys218Gly) and three with PLS3 c.73-24T>A, a published designation that does not correspond to an exon boundary in the MANE transcript NM_005032.7; five with previous bisphosphonates) received teriparatide 20 µg/day for 24 months: formation and resorption markers rose in the first 6 months (PINP +90% to +398%, osteocalcin +50% to +280%, CTX +58% to +457%), lumbar BMD rose by 5.2% to 7.9% in five patients and femoral neck BMD by 2.6% to 7.8% in four, distal radius cortical volumetric density fell by 5.4% to 26.1%, and biopsy showed increased osteoid (more consistent with WNT1), reduced eroded surface and increased marrow adipocytes [92]. Matrix mineralisation changed little on biopsy and marrow adiposity increased, and a man with the common LRP5 variant p.Arg1036Gln showed accelerated turnover on teriparatide [5, 37]. In pregnancy- and lactation-associated osteoporosis, a retrospective series of 52 women showed BMD gain without treatment, greater with bisphosphonates and greater still with teriparatide, with about 19% new fractures over 36 months in all groups [37].

7.6 Romosozumab and anti-sclerostin antibodies

Mechanism. Romosozumab is a humanised IgG2 monoclonal antibody that binds sclerostin and prevents LRP5/LRP6-mediated inhibition of the WNT pathway, increasing formation and, to a lesser extent, reducing resorption [25]. The effect in the forms of this class is an open question: if the LRP5 co-receptor is absent, activation depends on LRP6, and if the WNT1 ligand is insufficient, removing the brake may not restore the signal [39, 93]. In the mouse with Lrp5 inactivation, sclerostin inhibition retains an anabolic effect, but in the mouse with combined Lrp5 and Lrp6 inactivation it has no effect; in the Wnt1 “Swaying” mouse, anti-sclerostin antibody increased bone mass and reduced peripheral fractures; in the Pls3−/− mouse crossed with an Lrp5 high bone mass allele (mimicking anti-sclerostin therapy), cortical and trabecular mass increased [37, 39, 65]. These data are preclinical.

Dose and label. The approved dose is 210 mg subcutaneously (two 105 mg injections) once a month, for 12 months, administered by a healthcare professional, with calcium and vitamin D supplementation, followed by an antiresorptive agent if treatment continues to be justified [94, 95]. The approved indication in the USA (9 April 2019) is postmenopausal osteoporosis at high risk of fracture; in the EU (9 December 2019), severe postmenopausal osteoporosis at high risk [94, 95]. The US label carries a boxed warning for a possible increased risk of myocardial infarction, stroke and cardiovascular death, and the drug should not be started if there has been a myocardial infarction or stroke in the previous year [94]. Paediatric safety and efficacy have not been established [94]. Use in men, in premenopausal women and in any monogenic form is off-label.

Clinical evidence in the class. In a retrospective Hamburg cohort (2022–2024), 27 postmenopausal women with heterozygous LRP5, LRP6 or WNT1 variants (33 patient-years) were grouped according to treatment: no treatment beyond calcium and vitamin D (n=12), antiresorptive (n=11; IV ibandronate 3 mg every 3 months or denosumab 60 mg every 6 months) or romosozumab (n=10); the change in lumbar BMD at 12 months was −0.12 ± 5.01%, +3.97 ± 4.57% and +15.74 ± 7.63% (p<0.001), and that at the hip +0.41 ± 2.35%, +1.76 ± 2.11% and +5.24 ± 4.52%; formation markers rose more with romosozumab at 6 months and there were no new fractures in the romosozumab group over 1 year [39]. The authors acknowledge limitations: retrospective design, heterogeneous variants (including ACMG classes 2 and 3 and risk alleles), small groups, a densitometric outcome and lack of data in biallelic forms [39]. Case reports diverge. In a 67-year-old woman with a pathogenic LRP5 variant and four vertebral fractures, 1 year of romosozumab increased lumbar BMD by 58% (a value inflated by the fractures) and femoral neck BMD by 23%, but five new vertebral fractures occurred [96]. In a 67-year-old man with a novel WNT1 frameshift variant (NM_005430.4:c.578del p.(Asp193AlafsTer6), published as c.578delA) and recurrent fractures after 10 years of bisphosphonates and 6 of denosumab, romosozumab for 1 year was considered ineffective [93]. In a 28-year-old Brazilian woman with postpartum osteoporosis and a pathogenic WNT1 variant, after a partial response to teriparatide and denosumab, romosozumab contributed to a total gain of 42.1% in lumbar BMD and 16.6% at the hip [73]. In two postmenopausal women heterozygous for PLS3 variants with fractures, romosozumab “appeared effective”, according to an uncontrolled observation [65]. There are no data on romosozumab in OPPG or in children.

Setrusumab (investigational). Setrusumab is another anti-sclerostin antibody. In the phase IIb ASTEROID trial, in 90 adults with OI, the primary endpoint (trabecular volumetric density of the radius) was not met, but BMD increased in a dose-dependent manner [25]. On 29 December 2025, the sponsoring companies announced in a press release that the two phase 3 trials in OI had missed the primary endpoint, reduction in the annualised rate of clinical fractures: ORBIT (159 patients aged 5 to 25 years, 45 centres, 11 countries, 2:1 randomisation against placebo) and COSMIC (69 children aged 2 to under 7 years, 21 centres, 7 countries, 1:1 randomisation against IV bisphosphonates); both met the secondary BMD endpoint with statistical significance, the ORBIT placebo group had a low fracture rate, in COSMIC the reduction relative to bisphosphonates was not significant and the safety profile did not change; the press release did not disclose the numerical values [97]. This is a company press release result, with no scientific publication identified as of the date of this article, and it concerns OI and not the monogenic forms of this class.

7.7 Denosumab

Denosumab is an anti-RANKL antibody that inhibits osteoclast maturation; in adults, the approved dose is 60 mg SC every 6 months, with calcium 1000 mg and at least 400 IU of vitamin D per day, and pregnancy must be excluded before administration in women of childbearing potential [98]. The effect on turnover is rapidly reversible, and discontinuation may cause rebound resorption and severe hypercalcaemia in children, from 4 weeks after the injection; the first paediatric case of osteonecrosis of the jaw has been described in a giant cell tumour [75]. The US label mentions an open-label study in 153 children with OI (2–17 years) that did not demonstrate fracture reduction, hypercalcaemia requiring hospitalisation and acute kidney injury, and paediatric trials terminated early because of potentially fatal events; paediatric efficacy and safety have not been established [98]. GeneReviews does not recommend denosumab in children with OI [25]. In the Chinese WNT1 cohort, denosumab, like bisphosphonates, increased BMD [63]. In young adults, the risk of rebound with multiple vertebral fractures after discontinuation has not been studied, so the exit plan must be defined before starting [32, 37].

7.8 Therapeutic sequence in adults

In the two Finnish adults with WNT1 variants described by Ryhänen and colleagues, management required a sequence of drugs with long-term planning and an exit strategy, and the authors advocate access to anabolic agents for men and women with early-onset osteoporosis [32]. The logic is that of common osteoporosis: an anabolic agent (teriparatide or romosozumab) followed by an antiresorptive to consolidate the gain, as the romosozumab label recommends [75, 94]. In patients with low turnover, choosing an anabolic agent first-line has a biological rationale, but evidence of fracture reduction is lacking [5]. The TOPaZ trial, conducted at 27 centres between May 2017 and March 2025, randomised 350 adults with OI (mean age 43.7 years; 53.9% women; most with collagen-related OI type I) to teriparatide 20 µg/day for 2 years followed by zoledronic acid 5 mg IV, or usual care: incident fractures occurred in 65 of 176 (36.9%) versus 63 of 173 (36.4%), with an absolute risk reduction of −1.57% (95% CI −9.90% to 5.89%) and a hazard ratio of 0.97 (95% CI 0.68–1.38), despite a greater gain in lumbar and hip BMD and some favourable quality-of-life indices, with similar adverse events [99]. The authors concluded that bone quality weighs more than density in the genesis of fractures in OI, a conclusion that cautions against reading BMD gain as a surrogate for fracture reduction in this class too [99].

7.9 Regulatory table

7.10 Clinical studies in the class and limits of the evidence

None of these studies was designed to compare drugs with each other in the monogenic forms, and indirect comparisons between series are not valid, because age, gene, severity, previous treatments and measurement method differ. Almost all use BMD as the outcome; the two randomised trials with a fracture outcome available in 2025–2026, both in OI (TOPaZ and ORBIT/COSMIC), were negative for that outcome despite the densitometric gain [97, 99]. In children, BMD measured by DXA is confounded by growth and by bisphosphonate lines, and the positive results of the series do not demonstrate fracture reduction. The absence of fractures during short follow-up does not equate to zero risk.

7.11 Special situations

Pregnancy and lactation. In women with early-onset osteoporosis who wish to become pregnant, the timing of pregnancy, the effect of pregnancy and especially of lactation on bone mass and the timing of drugs are discussed [37]. A case-control study found no major teratogenic effects of bisphosphonates, but did not exclude effects on neonatal complications and on the live birth rate; as bisphosphonates are retained in bone, it is advised not to start them when a pregnancy is planned in the following year [37]. The APEG consensus recommends avoiding pregnancy in the 12 months following a bisphosphonate dose and performing a pregnancy test (urinary or serum hCG) before each dose in post-menarchal adolescents (strong recommendation, very low evidence), because of placental transfer and effects in animal models, although the few reports of exposure before conception have not shown significant fetal effects; in young people, renal excretion of pamidronate persists 8 years after discontinuation [78]. There are no human safety data for teriparatide, denosumab or romosozumab in pregnancy; as they are not retained in bone, they are assumed not to be teratogenic if stopped before conception, but the persistence of the effect without a subsequent bisphosphonate and the risk of rebound after denosumab are unknown [37]. LRP5 and WNT1 variants have been described in women with vertebral fractures in pregnancy or postpartum, which justifies genetic testing in these situations [37, 73].

Perioperative care and orthopaedics. Surgery for displaced fractures uses internal fixation and intramedullary rods [3]. After osteotomies, the bisphosphonate is withheld for a few months and only resumed with adequate callus [75]; in OI, APEG applies the same rule to fractures [78]. Prodromal pain in weight-bearing bones during prolonged treatment requires exclusion of atypical fracture [76].

Transition. The monogenic forms change with age: the fragility of SGMS2-related disease and of IJO decreases after puberty, whereas the vertebral compressions of WNT1-related osteoporosis arise mainly after the age of 50 and those of PLS3-related osteoporosis progress from childhood [9, 35, 49, 74]. Transition to adult medicine must maintain vertebral surveillance and decide on anabolic agents after epiphyseal closure [75, 91].

Agents to avoid. The sources do not list class-specific contraindicated drugs; glucocorticoids, anticonvulsants and cancer therapies are recognised as harmful to bone and should be weighed [37]. The APEG consensus contraindicates bisphosphonates in pregnancy, active rickets and disorders in which inhibiting resorption worsens fragility, such as hypophosphatasia and disorders with sclerotic lesions or high bone mass, a category that includes the gain-of-function pole of LRP5, and calls for extreme caution in renal failure [78]. The gain-of-function pole of LRP5 should not receive therapy aimed at the osteoporotic pole, so the direction of effect of the variant must be established before any decision [5].

8. Reproductive Options and Family Counselling

8.1 Recurrence risks by entity

OPPG (autosomal recessive). The parents of an affected individual are, as a rule, obligate heterozygotes, as confirmed in the Mennonite, Turkish, Greek and Brazilian families [38, 60, 69, 83]. In each pregnancy of a carrier couple, the risk is 25% for an affected child, 50% for a carrier and 25% for a non-carrier, according to recessive Mendelian transmission. Consanguinity is frequent in the published families (at least three consanguineous marriages in the Mennonite pedigree; consanguineous Turkish and Brazilian families) and increases the likelihood of carrier couples in endogamous communities [38, 69, 83]. In the extended Greek pedigree, among 44 relatives tested, 10 were homozygous and most of the rest heterozygous [60]. Counselling of carriers is not limited to reproductive risk: heterozygotes have reduced BMD, and 4 of 6 Mennonite parents had osteopenia or osteoporosis, including a 22-year-old mother with a Z-score in the osteoporotic range without fractures [38, 46, 57]. Identified carriers therefore benefit from DXA and bone health advice. Eye disease in carriers is rare and subtle (2 of 10 parents examined), but angiography may reveal changes in carriers of LRP5 variants [27, 52].

LRP5-, WNT1- and SGMS2-related osteoporosis (autosomal dominant). Each child of a carrier has a 50% chance of inheriting the variant. Expressivity is highly variable and there are carriers without a phenotype, which prevents severity from being predicted from the genotype [5, 8, 29]. In SGMS2-related disease, the p.(Arg50Ter) variant has been found in families with affected members over several generations and in apparently healthy carriers [8, 49]. In the biallelic WNT1 form, heterozygous parents have their own risk of dominant osteoporosis, which links the children's OI type XV to the parents' bone health [5, 42]. For LRP5, inheritance of the ocular forms may be dominant or recessive, and the same gene has gain-of-function variants causing high bone mass; variant classification shapes the counselling [5, 26].

PLS3-related osteoporosis (X-linked). A heterozygous woman transmits the variant to 50% of her sons, who will be hemizygous and usually affected, and to 50% of her daughters, who will be heterozygous with a variable phenotype, from normal BMD to osteoporosis with fractures in adulthood; an affected man transmits the variant to all his daughters and none of his sons, according to X-linked transmission [6, 37]. De novo variants exist, as in the Finnish girl with a Z-score of −6.6 at 6 years of age [36].

8.2 Cascade testing and minors

Identification of the variant in the proband allows testing of relatives; the Finnish authors recommend testing the children and close relatives of patients with WNT1-related osteoporosis for preventive measures, because the disease compromises bone mass accrual already in childhood and puberty [31]. Identification of a carrier relative justifies BMD assessment, optimisation of calcium, vitamin D and physical activity and, if necessary, treatment [37]. In the forms with manifestations in childhood, testing of minors has a direct clinical benefit (fracture prevention, vertebral surveillance, ophthalmological examination when LRP5 is suspected), which distinguishes it from predictive testing for disorders without intervention in childhood. Assessment of the siblings of a child with OPPG should include ophthalmological and radiological examination, because the bone disease may be silent in the first years [27]. In LRP5-related FEVR, angiography of apparently unaffected relatives reveals the true extent of family involvement [34, 52].

8.3 Prenatal and preimplantation diagnosis

With the familial variants identified, prenatal diagnosis by chorionic villus sampling or amniocentesis and preimplantation genetic diagnosis are possible [51]. The sources consulted do not describe prenatal ultrasound markers of OPPG; in SGMS2-related disease, femoral bowing has been seen on prenatal ultrasound [49]. The occurrence of variants of uncertain significance, especially in LRP5, argues against reproductive decisions based on variants not classified as pathogenic or likely pathogenic [28].

8.4 Pregnancy in a woman with monogenic osteoporosis

Pregnancy and lactation transiently reduce bone mass, and there are reports of postpartum vertebral fractures in women with LRP5 and WNT1 variants [37, 73]. Planning includes the timing of drugs (avoiding starting bisphosphonates in the year before a planned pregnancy and, according to APEG, avoiding pregnancy in the 12 months following a dose) and spinal surveillance during lactation [37]. In a woman with OPPG, blindness requires adapted support during pregnancy and in caring for the newborn.

9. Special Considerations: Population Diversity and Epidemiology

9.1 Frequency

OPPG. Orphanet classifies point prevalence in Europe as less than 1/1,000,000, with a mean value of 0.05/100,000 (about 1 in 2 million), an internal estimate not yet validated [20]. Ai and colleagues started from an estimated incidence of 1/2,000,000 and a carrier frequency of 1/700 to predict about 380,000 carriers in the USA; this is a calculation assumption and not a population measurement [27]. By 2008, about 50 cases had been published [38]. The disease affects several populations, with clusters of cases in consanguineous communities: conservative Pennsylvania Mennonites (homozygosity for NM_002335.4:c.1275G>A p.(Trp425Ter)), an extended Greek pedigree (NM_002335.4:c.2409_2503+79del), and Turkish, Israeli and Brazilian families [38, 57, 60, 69, 83]. Frontali and colleagues noted, in 1985, more families from Mediterranean countries [44]. The series of Ai included patients from several European countries, the Americas and Oceania, including São Paulo [27].

Monoallelic and X-linked forms. No population prevalence has been published for LRP5-related osteoporosis; Orphanet records three cases in the entity of LRP5-related primary osteoporosis (the Hartikka series) and unknown prevalence [21, 30]. Proportions in clinical cohorts measure diagnostic yield and not population frequency: relevant LRP5/LRP6 variants in 8.3% of 372 German adults with early-onset osteoporosis [28]; LRP5 variants in 3 of 20 Canadian children with primary osteoporosis [30]; PLS3 variants in 2 of 31 Finnish children with severe primary osteoporosis and in none of 64 with multiple fractures [36]; no pathogenic SGMS2 variant in 44 Finnish adults with idiopathic early-onset osteoporosis [80]; and 3 pathogenic or likely pathogenic variants in 37 Brazilian patients with idiopathic, severe or familial osteoporosis [64]. PLS3-related osteoporosis has a prevalence of less than 1/1,000,000, with 8 families in the Orphanet database and a Dutch cohort of 11 families [6, 7]. SGMS2-related disease had 26 patients from 11 families described up to 2023 [29]. Dominant WNT1-related osteoporosis has fewer than 20 families published up to 2022, two of them large Finnish families with the same variant [5]. For comparison, the worldwide point prevalence of OI of all causes is 8.06/100,000, with birth prevalences of 7.4/100,000 in Sweden, 6.95 in the USA, 6.4 in France and 1.5 in Ireland [22].

FEVR. In ocular screening of about 200,000 newborns, the incidence of FEVR was 0.11%, and a 2022 meta-analysis of more than 3200 patients attributed 13.6% of cases to LRP5, the largest fraction among the known genes [34]. In the historical GeneReviews text, LRP5 explained 18% (6/32) of cases of autosomal dominant FEVR in the original study and 10–25% in others [52].

9.2 Populations and regional data

The available data come mainly from Europe (Finland, the Netherlands, Germany, France, the United Kingdom, Greece), the USA (Pennsylvania Mennonites, Cleveland, Providence) and Canada [6, 27, 28, 30, 31, 34, 38, 60, 62]. The Finnish WNT1 variant p.(Cys218Gly) and the Mennonite LRP5 variant NM_002335.4:c.1275G>A p.(Trp425Ter) behave as founder variants in their respective communities [37, 38]. From Brazil, the sources consulted include the siblings with OPPG and the NP_002326.2:p.(Asn531Ile) variant and the young man treated with teriparatide (Federal University of São Paulo), the cohort of 37 patients with idiopathic or familial osteoporosis from the University of São Paulo and a woman with postpartum WNT1-related osteoporosis [64, 73, 83, 85]. The sources consulted do not include data from Portugal. The lack of data in the sources consulted does not demonstrate the absence of cases.

9.3 Special populations

Blind children with OPPG need adapted visual rehabilitation and fall-prevention programmes, and pain assessment depends on appropriate communication; cognitive impairment affects a minority [27]. Young women with early-onset osteoporosis require reproductive planning (8.4) [37]. Men with PLS3- or WNT1-related osteoporosis do not have approved access to romosozumab [94, 95].

9.4 European networks and patient associations

The European Reference Network on Rare Bone Diseases (ERN BOND) brings together centres with experience in genetic bone fragility; ERN-EYE covers rare eye diseases, such as the vitreoretinopathies [40, 41]. The ERN on rare endocrine conditions (ERN ENDO) and the research consortia GEFOS, GENOMOS and the COST action GEMSTONE study the genetics of osteoporosis [37]. Patient associations include the Portuguese Osteogenesis Imperfecta Association (APOI), the Osteogenesis Imperfecta Foundation in the USA and the Brittle Bone Society in the United Kingdom, which also support related bone fragility disorders [70, 100, 101].

10. Illustrative Clinical Case

10.1 Published case: siblings with a heterozygous LRP5 variant, fractures and subclinical vitreoretinopathy

Littman and colleagues, from Brown University (Providence, USA), described in 2023 two siblings referred to an osteoporosis and bone health clinic for successive fractures and growth delay [34]. The boy, aged 13, presented with a closed distal radius fracture sustained while playing sport; the history revealed at least three previous appendicular fractures (of both radial growth plates at different ages and of a metacarpal), all traumatic. Height was on the 8th centile and weight on the 4th centile; gait, spine and lumbar mobility were normal, without vertebral fractures, the elbows were hyperextensible, and sclerae and hearing were normal. The radiological skeletal survey showed widened medullary canals with thin cortices, gracile bones and coxa valga with a neck-shaft angle of about 150° (figure 1: https://pmc.ncbi.nlm.nih.gov/articles/PMC10040166/figure/f0001/), and DXA a Z-score of −3.0 (1st centile) (figure 2: https://pmc.ncbi.nlm.nih.gov/articles/PMC10040166/figure/f0002/) [34].

His sister, aged 10, had a similar history: fractures of a distal phalanx, of both forearm bones requiring surgical stabilisation and of the proximal humerus; height on the 68th centile, normal gait and spine, hyperextensible elbows, normal sclerae. Radiographs showed gracile bones, widened canals and thin cortices (figure 3: https://pmc.ncbi.nlm.nih.gov/articles/PMC10040166/figure/f0003/); the orthopantomogram revealed absence of five permanent teeth (upper lateral incisors, upper second premolars and lower right first premolar), which the authors described as altered dentinogenesis (figure 4: https://pmc.ncbi.nlm.nih.gov/articles/PMC10040166/figure/f0004/); DXA was on the 1st centile [34].

Exome sequencing identified in both the heterozygous variant NM_002335.4:c.1481G>A, NP_002326.2:p.(Arg494Gln), classified as likely pathogenic [34, 50]. As LRP5 affects bone and eye, ophthalmological assessment was requested: fluorescein angiography showed, in the boy, capillary loss in the temporal retina and a retinal hole in the right eye, probably atrophic due to peripheral ischaemia (figure 5: https://pmc.ncbi.nlm.nih.gov/articles/PMC10040166/figure/f0005/), and, in his sister, temporal capillary loss in the left eye; both were diagnosed with FEVR [34]. The boy underwent prophylactic 360° laser retinopexy in the right eye (figure 6: https://pmc.ncbi.nlm.nih.gov/articles/PMC10040166/figure/f0006/). At 2 years of joint orthopaedic and ophthalmological follow-up, vision was preserved in both, they had given up contact sports and there were no new fractures; BMD increased slowly, remaining below normal (Z-score of −2.6 in the boy) [34]. The team discussed pharmacological therapy with the parents and deferred it, because the fractures had been traumatic, there were no vertebral fractures and both were entering puberty; calcium and vitamin D supplementation was continued, with reassessment in the event of new fractures or insufficient bone gain [34].

10.2 Clinical reading of the case

The case illustrates three points of this article. First, the heterozygous LRP5 phenotype may reach the clinic through orthopaedics, with appendicular fractures that appear traumatic, and only targeted ophthalmological examination, with angiography, reveals the vitreoretinopathy, at a stage when laser can preserve vision [27, 34]. Second, the decision not to treat immediately followed the criteria that weigh fracture type, the spine and remaining growth, and not the isolated BMD value [75]; 2 years of follow-up without fractures does not allow conclusions about long-term risk. Third, the dental agenesis was interpreted by the authors as altered dentinogenesis, a term that in OI designates a dentine abnormality and not absence of teeth; the finding should not be used as a criterion for OI.

10.3 Radiological illustration of OPPG

Spranger's atlas documents a boy blind from birth: at 15 years of age, CT showed retrolental calcification of the persistent fetal retinal artery and primary vitreous, and spinal radiography showed generalised osteoporosis with anterior wedging of the vertebrae and lumbar kyphosis; at 18 years, a displaced fracture of the distal femoral shaft, with the dense proximal tibial metaphysis of bisphosphonate treatment; at 19 years, fixation with screws and an intramedullary rod, with severe osteoporosis and thin tibia and fibula [3].

11. Prognosis and Long-Term Follow-Up

11.1 Prognosis

OPPG. Blindness is congenital or early and progresses to total loss in most patients by about 15 years of age and in all by 25 [27]. Without treatment, the bone disease progresses and many patients become wheelchair-dependent in the second decade [3]. With bisphosphonates started early, the published series show marked improvement in BMD, less pain, vertebral reshaping and fewer fractures, although fractures persist in some patients and the response wanes in adults [38, 60, 84, 85]. An atypical femoral fracture has been described at 38 years of age [61]. No survival data have been published in the sources consulted.

Monoallelic LRP5-related osteoporosis. The course is heterogeneous, with fractures in three quarters of the carriers in the German cohort and persistently low bone formation [28].

WNT1-related osteoporosis. The disease worsens with age: vertebral compressions are rare before 50 years of age and affect 78% of older carriers, with kyphosis and height loss, and conventional treatment did not prevent vertebral pathology in several adults [35, 37].

PLS3-related osteoporosis. Men have progressive vertebral disease from childhood, with many fractures over their lifetime (median of 12 in the Dutch cohort) and BMD gain with prolonged treatment; women usually have a more benign course, with fractures mainly after the menopause [6, 65, 74].

SGMS2-related disease. Fragility decreases after puberty in the nonsense form; the missense form is severe from the neonatal period; the cranial nerve palsies remit spontaneously [5, 33, 49].

IJO. Most recover within 2 to 4 years after puberty; a minority retain deformities and disability [9].

11.2 Surveillance

There is no validated surveillance schedule for these disorders; the table brings together the measures proposed by the sources, with their respective origins.

In collagen-related OI, GeneReviews proposes orthopaedic assessment every 3 months up to 1 year of age, every 6 months from 1 to 3 years and then annually or after each new fracture; this schedule serves only as a reference for the severe early-onset forms of this class [25].

12. Summary and Final Considerations

The non-collagen monogenic primary osteoporoses form a class defined by bone fragility with onset in childhood or young adulthood, with low BMD and vertebral and peripheral fractures, and by mechanisms ranging from canonical WNT signalling (LRP5, WNT1, LRP6) to the actin cytoskeleton (PLS3) and sphingomyelin metabolism (SGMS2). Osteoporosis-pseudoglioma syndrome, recessive and caused by biallelic loss of function of LRP5, is the founding entity: congenital blindness due to failure of retinal vascularisation and severe childhood osteoporosis, with cognitive impairment in a minority and carriers with reduced bone mass. Monoallelic LRP5 and WNT1 variants cause dominant osteoporosis of variable expressivity; biallelic WNT1 variants cause a recessive OI; PLS3 mainly affects men, with vertebral disease from childhood; SGMS2 combines fractures, sclerotic calvarial lesions and transient cranial nerve palsies.

Diagnosis rests on the definition of osteoporosis in children and young adults, the exclusion of secondary causes, targeted ophthalmological examination and a bone fragility panel with copy-number analysis, or an exome or genome, with family segregation. The sensitivity of sequencing is limited by deletions, splicing variants and intronic variants, VUS are frequent in LRP5, and the same gene has a high bone mass pole, so the direction of effect is part of the interpretation. Idiopathic juvenile osteoporosis is only accepted after the monogenic forms have been excluded.

No drug has an approved indication for these disorders. Bisphosphonates, at weight-based paediatric doses extrapolated from OI according to the APEG consensus, with annual readjustment by DXA and discontinuation at the end of growth, improve BMD, pain and vertebral shape in OPPG and other forms, especially when started early, without proof of fracture reduction and with the risk of atypical fracture during long treatment. In adults, teriparatide and romosozumab increase BMD in small series of patients with LRP5, LRP6, WNT1 and PLS3 variants, with discordant responses in case reports; the randomised trials with a fracture outcome in 2025–2026, both in OI (teriparatide followed by zoledronate, and setrusumab), were negative despite BMD gain. Management is personalised and multidisciplinary, with vertebral surveillance, ophthalmological surveillance in LRP5-related forms, reproductive planning and cascade testing of relatives, within the ERN BOND and ERN-EYE networks.

13. References

1. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 603506: Low density lipoprotein receptor-related protein 5; LRP5. Available from: https://www.omim.org/entry/603506
2. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 14.1 Osteogenesis imperfecta; p. 497–500.
3. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 14.12 Osteoporosis-pseudoglioma syndrome; p. 540–542.
4. Unger S, Ferreira CR, Mortier GR, Ali H, Bertola DR, Calder A, et al. Nosology of genetic skeletal disorders: 2023 revision. Am J Med Genet A. 2023;191(5):1164-1209. doi:10.1002/ajmg.a.63132. PMID: 36779427.
5. Costantini A, Mäkitie RE, Hartmann MA, Fratzl-Zelman N, Zillikens MC, Kornak U, et al. Early-Onset Osteoporosis: Rare Monogenic Forms Elucidate the Complexity of Disease Pathogenesis Beyond Type I Collagen. J Bone Miner Res. 2022;37(9):1623-1641. doi:10.1002/jbmr.4668. PMID: 35949115.
6. Zervou Z, Bruggenwirth HT, Demirdas S, Zillikens MC. Phenotypic and genetic characteristics of a Dutch cohort of patients with X-linked osteoporosis due to PLS3 genetic variants. JBMR Plus. 2025;9(6):ziaf046. doi:10.1093/jbmrpl/ziaf046. PMID: 40353206.
7. Orphanet. X-linked osteoporosis with fractures (ORPHA:391330). Available from: https://www.orpha.net/en/disease/detail/391330
8. Robinson ME, Bardai G, Veilleux LN, Glorieux FH, Rauch F. Musculoskeletal phenotype in two unrelated individuals with a recurrent nonsense variant in SGMS2. Bone. 2020;134:115261. doi:10.1016/j.bone.2020.115261. PMID: 32028018.
9. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 14.8 Idiopathic juvenile osteoporosis; p. 526–528.
10. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 259770: Osteoporosis-pseudoglioma syndrome; OPPG. Available from: https://www.omim.org/entry/259770
11. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 166710: Osteoporosis. Available from: https://www.omim.org/entry/166710
12. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 615221: Bone mineral density quantitative trait locus 16; BMND16 (osteoporosis, early-onset, autosomal dominant). Available from: https://www.omim.org/entry/615221
13. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 615220: Osteogenesis imperfecta, type XV; OI15. Available from: https://www.omim.org/entry/615220
14. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 164820: Wingless-type MMTV integration site family, member 1; WNT1. Available from: https://www.omim.org/entry/164820
15. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 300910: Bone mineral density quantitative trait locus 18; BMND18. Available from: https://www.omim.org/entry/300910
16. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 300131: Plastin 3; PLS3. Available from: https://www.omim.org/entry/300131
17. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 126550: Calvarial doughnut lesions with bone fragility; CDL. Available from: https://www.omim.org/entry/126550
18. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 611574: Sphingomyelin synthase 2; SGMS2. Available from: https://www.omim.org/entry/611574
19. Online Mendelian Inheritance in Man, OMIM®. Johns Hopkins University, Baltimore, MD. MIM Number: 259750: Osteoporosis, juvenile. Available from: https://www.omim.org/entry/259750
20. Orphanet. Osteoporosis-pseudoglioma syndrome (ORPHA:2788). Available from: https://www.orpha.net/en/disease/detail/2788
21. Orphanet. LRP5-related primary osteoporosis (ORPHA:498481). Available from: https://www.orpha.net/en/disease/detail/498481
22. Orphanet. Osteogenesis imperfecta (ORPHA:666). Available from: https://www.orpha.net/en/disease/detail/666
23. Orphanet. Calvarial doughnut lesions-bone fragility syndrome (ORPHA:85192). Available from: https://www.orpha.net/en/disease/detail/85192
24. Orphanet. Idiopathic juvenile osteoporosis (ORPHA:85193). Available from: https://www.orpha.net/en/disease/detail/85193
25. Rodriguez Celin M, Steiner RD, Basel D. COL1A1- and COL1A2-Related Osteogenesis Imperfecta. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 2005 Jan 28 [updated 2025 May 29]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1295/
26. Scruggs BA, Schimmenti LA. NDP-Related Retinopathies. In: Adam MP, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 1999 Jul 30 [revised 2023 Mar 23]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1331/ PMID: 20301506.
27. Ai M, Heeger S, Bartels CF, Schelling DK. Clinical and molecular findings in osteoporosis-pseudoglioma syndrome. Am J Hum Genet. 2005;77(5):741-753. doi:10.1086/497706. PMID: 16252235.
28. Stürznickel J, Rolvien T, Delsmann A, Butscheidt S, Barvencik F, Mundlos S, et al. Clinical Phenotype and Relevance of LRP5 and LRP6 Variants in Patients With Early-Onset Osteoporosis (EOOP). J Bone Miner Res. 2021;36(2):271-282. doi:10.1002/jbmr.4197. PMID: 33118644.
29. Merkuryeva E, Markova T, Tyurin A, Valeeva D, Kenis V, Sumina M, et al. Clinical and Genetic Characteristics of Calvarial Doughnut Lesions with Bone Fragility in Three Families with a Reccurent SGMS2 Gene Variant. Int J Mol Sci. 2023;24(9):8021. doi:10.3390/ijms24098021. PMID: 37175737.
30. Hartikka H, Mäkitie O, Männikkö M, Doria AS, Daneman A, Cole WG, et al. Heterozygous mutations in the LDL receptor-related protein 5 (LRP5) gene are associated with primary osteoporosis in children. J Bone Miner Res. 2005;20(5):783-789. doi:10.1359/jbmr.050101. PMID: 15824851.
31. Mäkitie RE, Haanpää M, Valta H, Pekkinen M, Laine CM, Lehesjoki AE, et al. Skeletal Characteristics of WNT1 Osteoporosis in Children and Young Adults. J Bone Miner Res. 2016;31(9):1734-1742. doi:10.1002/jbmr.2841. PMID: 27005318.
32. Ryhänen EM, Mäkitie RE, Pekkarinen T, Kröger H, Tong X, Kerttula L, et al. Severe early-onset osteoporosis due to heterozygous WNT1 variants in adults: a clinical and therapeutic challenge. J Bone Miner Res. 2025;41(1):85-90. doi:10.1093/jbmr/zjaf150. PMID: 41128774.
33. Pekkinen M, Terhal PA, Botto LD, Henning P, Mäkitie RE, Roschger P, et al. Osteoporosis and skeletal dysplasia caused by pathogenic variants in SGMS2. JCI Insight. 2019;4(7):126180. doi:10.1172/jci.insight.126180. PMID: 30779713.
34. Littman J, Phornphutkul C, Saade C, Katarincic J, Aaron R. Osteoporosis, Fractures, and Blindness Due to a Missense Mutation in the LRP5 Receptor. Orthop Res Rev. 2023;15:39-45. doi:10.2147/orr.s400111. PMID: 36987449.
35. Mäkitie RE, Niinimäki T, Nieminen MT, Schalin-Jäntti C, Niinimäki J, Mäkitie O. Impaired WNT signaling and the spine-Heterozygous WNT1 mutation causes severe age-related spinal pathology. Bone. 2017;101:3-9. doi:10.1016/j.bone.2017.04.001. PMID: 28411110.
36. Kämpe AJ, Costantini A, Mäkitie RE, Jäntti N, Valta H, Mäyränpää M, et al. PLS3 sequencing in childhood-onset primary osteoporosis identifies two novel disease-causing variants. Osteoporos Int. 2017;28(10):3023-3032. doi:10.1007/s00198-017-4150-9. PMID: 28748388.
37. Mäkitie O, Zillikens MC. Early-Onset Osteoporosis. Calcif Tissue Int. 2022;110(5):546-561. doi:10.1007/s00223-021-00885-6. PMID: 34236445.
38. Streeten EA, McBride D, Puffenberger E, Hoffman ME, Pollin TI, Donnelly P, et al. Osteoporosis-pseudoglioma syndrome: description of 9 new cases and beneficial response to bisphosphonates. Bone. 2008;43(3):584-590. doi:10.1016/j.bone.2008.04.020. PMID: 18602879.
39. von Brackel FN, Dehne N, Kornak U, Barvencik F, Schinke T, Amling M, et al. Romosozumab is effective in patients with variants in LRP5, LRP6, or WNT1. Osteoporos Int. 2026;37(5):1267-1278. doi:10.1007/s00198-026-07932-8. PMID: 41813941.
40. ERN BOND — European Reference Network on Rare Bone Diseases. Available from: https://ernbond.eu/
41. ERN-EYE — European Reference Network for Rare Eye Diseases. Available from: https://www.ern-eye.eu/
42. Laine CM, Joeng KS, Campeau PM, Kiviranta R, Tarkkonen K, Grover M, et al. WNT1 mutations in early-onset osteoporosis and osteogenesis imperfecta. N Engl J Med. 2013;368(19):1809-1816. doi:10.1056/nejmoa1215458. PMID: 23656646.
43. van Dijk FS, Zillikens MC, Micha D, Riessland M, Marcelis CL, de Die-Smulders CE, et al. PLS3 mutations in X-linked osteoporosis with fractures. N Engl J Med. 2013;369(16):1529-1536. doi:10.1056/nejmoa1308223. PMID: 24088043.
44. Frontali M, Stomeo C, Dallapiccola B. Osteoporosis-pseudoglioma syndrome: report of three affected sibs and an overview. Am J Med Genet. 1985;22(1):35-47. doi:10.1002/ajmg.1320220104. PMID: 3931475.
45. Gong Y, Vikkula M, Boon L, Liu J, Beighton P, Ramesar R, et al. Osteoporosis-pseudoglioma syndrome, a disorder affecting skeletal strength and vision, is assigned to chromosome region 11q12-13. Am J Hum Genet. 1996;59(1):146-151. PMCID: PMC1915094. PMID: 8659519.
46. Gong Y, Slee RB, Fukai N, Rawadi G, Roman-Roman S, Reginato AM, et al. LDL receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell. 2001;107(4):513-523. doi:10.1016/s0092-8674(01)00571-2. PMID: 11719191.
47. Keupp K, Beleggia F, Kayserili H, Barnes AM, Steiner M, Semler O, et al. Mutations in WNT1 cause different forms of bone fragility. Am J Hum Genet. 2013;92(4):565-574. doi:10.1016/j.ajhg.2013.02.010. PMID: 23499309.
48. Pyott SM, Tran TT, Leistritz DF, Pepin MG, Mendelsohn NJ, Temme RT, et al. WNT1 mutations in families affected by moderately severe and progressive recessive osteogenesis imperfecta. Am J Hum Genet. 2013;92(4):590-597. doi:10.1016/j.ajhg.2013.02.009. PMID: 23499310.
49. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 14.15 Calvarial doughnut lesions-osteoporosis syndrome; p. 548–551.
50. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-8. doi:10.1002/humu.23348. Available from: https://variantvalidator.org/
51. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Retinal dysplasia; p. 300–301.
52. Toomes C, Downey L. Familial Exudative Vitreoretinopathy, Autosomal Dominant – Retired chapter, for historical reference only. In: Adam MP, Feldman J, Mirzaa GM, et al., editors. GeneReviews®. Seattle (WA): University of Washington, Seattle; 2005 Mar 21 [updated 2011 Jul 14]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1147/ PMID: 20301326.
53. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Increased bone density; p. 198–199.
54. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 16.1 Osteopetroses; p. 575–576.
55. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 16.5 Osteopetrosis, late onset forms (MIM 607634, 166660); p. 587–590.
56. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 17.5 Endosteal hyperostosis, van Buchem type (MIM 239100, 269500); p. 634–638.
57. Lev D, Binson I, Foldes AJ, Watemberg N, Lerman-Sagie T. Decreased bone density in carriers and patients of an Israeli family with the osteoporosis-pseudoglioma syndrome. Isr Med Assoc J. 2003;5(6):419-421. PMID: 12841014.
58. Laine CM, Chung BD, Susic M, Prescott T, Semler O, Fiskerstrand T, et al. Novel mutations affecting LRP5 splicing in patients with osteoporosis-pseudoglioma syndrome (OPPG). Eur J Hum Genet. 2011;19(8):875-881. doi:10.1038/ejhg.2011.42. PMID: 21407258.
59. Chung BD, Kayserili H, Ai M, Freudenberg J, Uzümcü A, Uyguner O, et al. A mutation in the signal sequence of LRP5 in a family with an osteoporosis-pseudoglioma syndrome (OPPG)-like phenotype indicates a novel disease mechanism for trinucleotide repeats. Hum Mutat. 2009;30(4):641-648. doi:10.1002/humu.20916. PMID: 19177549.
60. Papadopoulos I, Bountouvi E, Attilakos A, Gole E, Dinopoulos A, Peppa M, et al. Osteoporosis-pseudoglioma syndrome: clinical, genetic, and treatment-response study of 10 new cases in Greece. Eur J Pediatr. 2019;178(3):323-329. doi:10.1007/s00431-018-3299-3. PMID: 30499050.
61. Alonso N, Soares DC, V McCloskey E, Summers GD, Ralston SH, Gregson CL. Atypical femoral fracture in osteoporosis pseudoglioma syndrome associated with two novel compound heterozygous mutations in LRP5. J Bone Miner Res. 2015;30(4):615-620. doi:10.1002/jbmr.2403. PMID: 25384351.
62. Collet C, Ostertag A, Ricquebourg M, Delecourt M, Tueur G, Isidor B, et al. Primary Osteoporosis in Young Adults: Genetic Basis and Identification of Novel Variants in Causal Genes. JBMR Plus. 2018;2(1):12-21. doi:10.1002/jbm4.10020. PMID: 30283887.
63. Hu J, Lin X, Gao P, Zhang Q, Zhou B, Wang O, et al. Genotypic and Phenotypic Spectrum and Pathogenesis of WNT1 Variants in a Large Cohort of Patients With OI/Osteoporosis. J Clin Endocrinol Metab. 2023;108(7):1776-1786. doi:10.1210/clinem/dgac752. PMID: 36595228.
64. Rocha-Braz MGM, França MM, Fernandes AM, Lerario AM, Zanardo EA, de Santana LS, et al. Comprehensive Genetic Analysis of 128 Candidate Genes in a Cohort With Idiopathic, Severe, or Familial Osteoporosis. J Endocr Soc. 2020;4(12):bvaa148. doi:10.1210/jendso/bvaa148. PMID: 33195954.
65. Windels O, Simon A, Schiewek L, Delsmann J, Kornak U, Amling M, et al. Skeletal pathologies caused by PLS3 inactivation are predominantly associated with impaired cortical bone integrity. J Bone Miner Res. 2026 [Epub ahead of print]:zjag141. doi:10.1093/jbmr/zjag141. PMID: 42784573.
66. Pihlström S, Oghabian A, Määttä K, Legebeke J, Mäkitie RE, Campeau PM, et al. Transcriptomic and lipidomic profiling provide novel insight into the pathogenesis of monogenic SGMS2-related osteoporosis. JBMR Plus. 2025;9(10):ziaf128. doi:10.1093/jbmrpl/ziaf128. PMID: 40978119.
67. Patra S, Jena S, Kedar K, Pande M, Katam KK, Prajapti A, et al. A novel SGMS2 mutation associated with high bone mass; description of an affected family with recurrent fragility fractures. Bone Rep. 2025;24:101833. doi:10.1016/j.bonr.2025.101833. PMID: 40123745.
68. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 40, section 40.1 Disorders of Sphingolipid Synthesis (40.1.9 Sphingomyelin Synthase 2 Mutations and Osteoporosis).
69. Tüysüz B, Bursalı A, Alp Z, Suyugül N, Laine CM, Mäkitie O. Osteoporosis-pseudoglioma syndrome: three novel mutations in the LRP5 gene and response to bisphosphonate treatment. Horm Res Paediatr. 2012;77(2):115-120. doi:10.1159/000336193. PMID: 22456437.
70. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. Fractures; p. 158–160.
71. Balagamage C, Sen P, Crabtree N, Williams D, Uday S. LRP5-related primary osteoporosis: phenotypic spectrum and treatment response to zoledronic acid. J Pediatr Endocrinol Metab. 2026;39(5):499-507. doi:10.1515/jpem-2025-0652. PMID: 41843941.
72. Wang Q, Liu M, Cao BY, Su C, Meng X, Ding Y, et al. Osteoporosis Caused by Monoallelic Variant of WNT1 Gene in Four Pediatric Patients. Am J Med Genet A. 2025;197(5):e63987. doi:10.1002/ajmg.a.63987. PMID: 39780405.
73. Cavalcanti Matos T, Kistemann Gonçalves Dias M, Franco AS. Use of Teriparatide, Denosumab, and Romosozumab in a Postpartum Monogenic Osteoporosis With a WNT1 Pathogenic Variation. JCEM Case Rep. 2025;3(4):luaf053. doi:10.1210/jcemcr/luaf053. PMID: 40144478.
74. Mäkitie RE, Niinimäki T, Suo-Palosaari M, Kämpe A, Costantini A, Toiviainen-Salo S, et al. PLS3 Mutations Cause Severe Age and Sex-Related Spinal Pathology. Front Endocrinol (Lausanne). 2020;11:393. doi:10.3389/fendo.2020.00393. PMID: 32655496.
75. Sakka SD, Cheung MS. Management of primary and secondary osteoporosis in children. Ther Adv Musculoskelet Dis. 2020;12:1759720X20969262. doi:10.1177/1759720x20969262. PMID: 33224280.
76. van de Laarschot DM, Zillikens MC. Atypical femur fracture in an adolescent boy treated with bisphosphonates for X-linked osteoporosis based on PLS3 mutation. Bone. 2016;91:148-151. doi:10.1016/j.bone.2016.07.022. PMID: 27477003.
77. Bishop N, Arundel P, Clark E, Dimitri P, Farr J, Jones G, et al. Fracture prediction and the definition of osteoporosis in children and adolescents: the ISCD 2013 Pediatric Official Positions. J Clin Densitom. 2014;17(2):275-280. doi:10.1016/j.jocd.2014.01.004. PMID: 24631254.
78. Simm PJ, Biggin A, Zacharin MR, Rodda CP, Tham E, Siafarikas A, et al. Consensus guidelines on the use of bisphosphonate therapy in children and adolescents. J Paediatr Child Health. 2018;54(3):223-233. doi:10.1111/jpc.13768. PMID: 29504223.
79. Weber DR, Boyce A, Gordon C, Högler W, Kecskemethy HH, Misra M, et al. The Utility of DXA Assessment at the Forearm, Proximal Femur, and Lateral Distal Femur, and Vertebral Fracture Assessment in the Pediatric Population: 2019 ISCD Official Position. J Clin Densitom. 2019;22(4):567-589. doi:10.1016/j.jocd.2019.07.002. PMID: 31421951.
80. Loid P, Richardt S, Niinimäki T, Pekkinen M, Mäkitie O, Mäkitie R. Pathogenic SGMS2 variants are not a common cause of early-onset osteoporosis among Finnish patients. Front Endocrinol (Lausanne). 2026;17:1719375. doi:10.3389/fendo.2026.1719375. PMID: 41821748.
81. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 14.13 Spondyloocular dysplasia; p. 543–544.
82. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford: Oxford University Press; 2018. 14.2 Osteogenesis imperfecta, type I (MIM 116200); p. 501–505.
83. Barros ER, Dias da Silva MR, Kunii IS, Lazaretti-Castro M. Three years follow-up of pamidronate therapy in two brothers with osteoporosis-pseudoglioma syndrome (OPPG) carrying an LRP5 mutation. J Pediatr Endocrinol Metab. 2008;21(8):811-818. doi:10.1515/jpem.2008.21.8.811. PMID: 18825883.
84. Zacharin M, Cundy T. Osteoporosis pseudoglioma syndrome: treatment of spinal osteoporosis with intravenous bisphosphonates. J Pediatr. 2000;137(3):410-415. doi:10.1067/mpd.2000.107838. PMID: 10969269.
85. Arantes HP, Barros ER, Kunii I, Bilezikian JP, Lazaretti-Castro M. Teriparatide increases bone mineral density in a man with osteoporosis pseudoglioma. J Bone Miner Res. 2011;26(12):2823-2826. doi:10.1002/jbmr.530. PMID: 21997141.
86. Velandia-Avendaño MC, Sarmiento-Ramón MP. X-linked Osteoporosis due to PLS3 Pathogenic Variant: Case Report on Zoledronic Acid Treatment in Siblings. J Clin Res Pediatr Endocrinol. 2026;18(3):543-548. doi:10.4274/jcrpe.galenos.2025.2025-4-2. PMID: 40959584.
87. U.S. Food and Drug Administration. Pamidronate disodium injection — prescribing information (ANDA075841), revised 2026-05; Aredia NDA020036 approved 1991-10-31. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=020036
88. U.S. Food and Drug Administration. Zoledronic acid injection — prescribing information (ANDA201783), revised 2023-09; Reclast NDA021817 approved 2007-04-16. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=021817
89. U.S. Food and Drug Administration. BINOSTO (alendronate sodium) — prescribing information (NDA202344), revised 2026-02. Available from: https://api.fda.gov/drug/label.json?search=openfda.application_number:NDA202344
90. U.S. Food and Drug Administration. Teriparatide injection — prescribing information (ANDA213641), revised 2026-03; Forteo NDA021318 approved 2002-11-26. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=021318
91. European Medicines Agency. Forsteo (teriparatide) — EPAR. Authorised 2003-06-10. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/forsteo
92. Välimäki VV, Mäkitie O, Pereira R, Laine C, Wesseling-Perry K, Määttä J, et al. Teriparatide Treatment in Patients With WNT1 or PLS3 Mutation-Related Early-Onset Osteoporosis: A Pilot Study. J Clin Endocrinol Metab. 2017;102(2):535-544. doi:10.1210/jc.2016-2423. PMID: 27732335.
93. Bailey R, Gee C, Schoenhoff G, Dai-Ju JQ. A Case of Early-Onset Osteoporosis Due to a Novel WNT1 Variant. AACE Endocrinol Diabetes. 2026;13(2):352-355. doi:10.1016/j.aed.2025.12.003. PMID: 41938316.
94. U.S. Food and Drug Administration. EVENITY (romosozumab-aqqg) injection — prescribing information (BLA761062), revised 2026-07. DailyMed/openFDA. Available from: https://api.fda.gov/drug/label.json?search=openfda.application_number:BLA761062
95. European Medicines Agency. Evenity (romosozumab) — EPAR. Authorised 2019-12-09. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/evenity
96. van Velsen EFS, Wijnen M, Muradin GSR, Zillikens MC. Incident Vertebral Fractures During Romosozumab Treatment in a Patient With a Pathogenic LRP5 Variant. JCEM Case Rep. 2025;3(1):luae238. doi:10.1210/jcemcr/luae238. PMID: 39726666.
97. Mereo BioPharma. Mereo BioPharma announces Phase 3 Orbit and Cosmic results for setrusumab (UX143) in osteogenesis imperfecta [press release]. 2025 Dec 29. Available from: https://media.mereobiopharma.com/news/phase-3-orbit-and-cosmic-results/
98. U.S. Food and Drug Administration. Denosumab-bmwo (Stoboclo) — prescribing information (BLA761404), revised 2026-02; Prolia BLA125320 approved 2010-06-01. Available from: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=125320
99. Hald JD, Weir CJ, Keerie C, Dewar L, MacLean M, Milne L, et al. Teriparatide Plus Zoledronic Acid for Osteogenesis Imperfecta: A Randomized Clinical Trial. JAMA. 2026;336(2):116-124. doi:10.1001/jama.2026.6889. PMID: 42133304.
100. APOI — Associação Portuguesa de Osteogénese Imperfeita. Available from: https://www.apoi.pt/
101. Osteogenesis Imperfecta Foundation. Available from: https://oif.org/