Friend or foe: high bone mineral density on routine bone density scanning, a review of causes and management

Abstract

A finding of high BMD on routine DXA scanning is not infrequent and most commonly reflects degenerative disease. However, BMD increases may also arise secondary to a range of underlying disorders affecting the skeleton. Although low BMD increases fracture risk, the converse may not hold for high BMD, since elevated BMD may occur in conditions where fracture risk is increased, unaffected or reduced. Here we outline a classification for the causes of raised BMD, based on identification of focal or generalized BMD changes, and discuss an approach to guide appropriate investigation by clinicians after careful interpretation of DXA scan findings within the context of the clinical history. We will also review the mild skeletal dysplasia associated with the currently unexplained high bone mass phenotype and discuss recent advances in osteoporosis therapies arising from improved understanding of rare inherited high BMD disorders.

Definition of high BMD

BMD measurement plays an important role in the assessment of osteoporosis and fracture risk. In clinical practice, BMD is most commonly measured using DXA. BMD is then compared against an age-, ethnicity- and gender-specific reference population to compute T- and Z-scores [number of standard deviations a measured BMD differs from the mean BMD of a young adult population (T-score) or age-matched population (Z-score)]. In 1994 the World Health Organization defined osteoporosis in terms of BMD and fracture, a T-score of ≤−2.5 and/or a previous fragility fracture [1]. Equivalent definitions for high BMD do not currently exist. While low BMD relates to increased fracture risk, the converse may not hold for high BMD. As we will discuss, high BMD may occur in conditions (i) with increased fracture risk [e.g. osteopetrosis or Paget’s Disease (PD)] or (ii) such as artefacts that themselves do not affect fracture risk but may mask low BMD and (iii) where fracture risk may be reduced but other comorbidities may exist that are only starting to be recognized.

The absence of an upper limit for BMD may risk those with high BMD, potentially due to underlying pathology, being labelled as normal [2]. In 2005, Michael Whyte [2] advocated a high BMD definition as a Z-score >+2.5 to highlight to clinicians the potential for underlying pathology. Epidemiological studies of high BMD are few and definition thresholds variable [3, 4]. Until recently, high BMD was usually the reserve of case reports and case series. The first systematic analysis of patients undergoing routine clinical DXA scanning, encompassing 335 115 DXA scans across 15 UK centres, used a screening threshold T- or Z-score ≥+4 at any lumbar/hip site [5]. This study was the first to assess the prevalence of high BMD within the general population referred for DXA scanning.

Prevalence of high BMD

If BMD is normally distributed, then a threshold Z-score of ≥+2.5 should by definition identify 6.2/1000, and a more extreme Z-score ≥+4 would identify 3/100 000 [6]. In fact, based on assessment at 13 UK hospitals, 5/1000 NHS DXA scans have a T/Z-score ≥+4, approximately half of which are explained by artefactual elevations in BMD resulting from osteoarthritic degeneration. Of these incidental cases with high BMD, 35% had been referred due to a suspicion of osteoporosis and 22% because of an underlying medical condition necessitating bone assessment [5].

Causes of high BMD

While a finding of high BMD on conventional DXA scanning most commonly reflects degenerative disease, increases in BMD can also arise secondary to an underlying disorder with skeletal effects. Here we outline a classification for the causes of raised BMD seen on DXA scanning (summarized in Table 1).

Table 1.

Classification of the potential causes of a high BMD value detected by DXA scanning

Artefactual causes of raised BMD—no true increase in bone mass
OA
DISH
AS
Vertebral fractures
Vascular calcification
Thalassaemia major
Abdominal abscesses
Gallstones
Renal calculi
Gluteal silicon implants
Gaucher’s disease
Intestinal barium
Surgical metalwork
Laminectomy
Vertebroplasty and kyphoplasty

True causes of increased bone mass and density
Localized	Acquired	PDa
		Hypophosphatasiasa
		Tumours	Primary malignancies, e.g. osteoblastoma
			Secondary metastases, e.g. prostate
			Other tumours
		SAPHO syndrome
		Chronic infective osteomyelitis
		Osseous tuberous sclerosis
Generalized	Acquired	Fluorosis
		Renal osteodystrophy
		Acromegaly
		Hepatitis C-associated osteosclerosis
		Myelofibrosis
		Mastocytosis
		Oestrogen replacement implants
	Congenital	Reduced bone resorption (Table 2)	Osteopetrosis
			Pycnodysostosis
			Osteopoikilosis
			Melorheostosis
		Increased bone formation (Table 3)	Sclerosteosis
			van Buchem’s disease
			LRP5 HBM
			LRP4 HBM
			Craniometaphyseal dysplasia
		Disturbed formation and resorption (Table 3)	Camurati–Engelmann disease
			Ghosal syndrome
		Unexplained HBM

^aMay be congenital.

Artefactual elevations in BMD measurements

Artefactually raised BMD values do not equate to a true increase in bone mass, but usually result from artefactual elevations in calcium content, which can be recognized by careful inspection of the DXA scan in the context of the clinical history; some examples are shown in Fig. 1. Artefact is important to differentiate, as it may mask osteoporosis.

Fig. 1.

Examples of DXA images identified with a T/Z-score ≥+4.

(A) Artefactually raised lumbar spine BMD due to osteoarthritic spondylosis. (B) Artefactually raised lumbar spine BMD due to ankylosing spondylitis; anterior longitudinal ligament ossification is seen. (C) Generalized increase in lumbar spine BMD at all vertebral levels in a case of unexplained HBM.

Osteoarthritic spondylosis

Osteoarthritic spondylosis most commonly explains artefactual elevations in calcium content due to abnormally dense bone at the vertebral margins forming vertebral end-plate sclerosis, facet joint sclerosis and osteophytes (Fig. 1A). Facet joint OA is particularly marked in the lower lumbar spine, giving the recognized pattern of progressive osteoarthritic changes seen in sequential descending lumbar vertebrae, which correlates with rising BMD measures caudally down the spine [7]. Even mild osteophytosis can result in a 24% increase in lumbar BMD [8]. Osteoarthritic spondylosis accounts for 49% of T/Z-score ≥+4 on routine DXA assessments [5]. Conversely, osteoarthritic effects on femoral neck BMD are minimal [9].

In clinical practice, where osteoarthritic changes are restricted to one or two vertebrae, these are excluded and the lumbar spine DXA result is based on the mean value of unaffected vertebrae. Confirmatory radiographs are generally not required, as changes suggestive of spondylosis (e.g. end-plate sclerosis, preferential effects on lower lumbar vertebrae) are evident on DXA scan inspection, which may also reveal abnormalities underlying osteoarthritic changes (e.g. scoliosis).

Diffuse idiopathic skeletal hyperostosis

Diffuse idiopathic skeletal hyperostosis (DISH) is a skeletal disorder characterized by widespread calcification at spinal and extra-spinal sites. Although the aetiology is unknown, DISH has been associated with features of the metabolic syndrome [10, 11]. Ossification of spinal ligaments in DISH can overestimate vertebral areal BMD from 24% to 39% and may mask osteoporosis on DXA scanning [12, 13]. Among older men, in whom DISH is common, DISH has been associated with increased vertebral fracture risk [14]. The prevalence of DISH rises sharply with age and varies according to ethnicity [15].

Ankylosing spondylitis

Syndesmophyte formation at vertebral margins in advanced AS can elevate spinal BMD by increasing calcium content [16]. This is compounded by anterior longitudinal ligament ossification, plus coexistent scoliosis and inflammation (Fig. 1B). Spinal DXA BMD measurements may therefore be high despite loss of trabecular bone resulting in increased fracture risk (particularly vertebral fracture) [17, 18]. Hip BMD is affected less by bony changes in AS and therefore hip DXA has been suggested as a more reliable method to assess fracture risk in these patients [17, 18].

Vertebral fracture

In vertebral fracture, bone mineral content is unchanged, but BMD increases due to a reduction in the denominator (i.e. vertebral area). Although absolute elevations in BMD may be modest, this mechanism is a common artefactual cause for BMD gain during serial DXA monitoring for osteoporosis [19]. Reduction in vertebral area contrasts with the normal finding of successive increases in vertebral area when moving down the spine. In clinical practice, affected vertebrae should be excluded from DXA analysis and mean BMD calculated from the remaining lumbar vertebrae. Although vertebral fractures can be detected by conventional lumbar DXA, vertebral height loss is more accurately quantified by lateral DXA [20]. Following vertebroplasty, polymethylmethacrylate cement will also elevate measured BMD.

Extrinsic artefacts

Calcification of structures anterior to the spine but within the DXA field can artefactually elevate BMD measurements. Although vascular calcification of the abdominal aorta is common, reported in 43% of patients having lumbar DXA assessment (mean age 68 years), there is little evidence from human studies that this significantly affects lumbar spine BMD measures [7, 8, 21–23]. Other radiodense materials can elevate BMD values. Soft tissue iron deposition in thalassaemia major, usually associated with osteoporosis, has been reported to lead to a T-score of up to +4.9 when, interestingly, the lateral DXA view showed the increased density to lie anterior to the vertebral body with the remaining vertebrae registering a T-score of +0.30, presumably representing soft tissue iron deposition [24]. Similarly, abdominal abscesses which can calcify [25], gallstones [26, 27], renal calculi [27] and gluteal silicon implants [28] have been linked to erroneously high BMD values. Gaucher’s disease, with excess glycolipid within an overlying enlarged spleen, has been associated with high BMD, particularly at L1 (Z-score +3.8), despite coexistent low hip BMD, possibly reflecting the high glycolipid load or secondary calcification in the spleen [5]. Radiological barium administration into overlying bowel may falsely elevate BMD, though this has not been reported to date. Surgical metalwork explains 1.4% of incidental high BMDs on routine DXA scanning [5]. Laminectomy can also increase BMD values [29].

Focal abnormalities causing increased BMD measurements

Focal increases in bone mass can significantly alter BMD measurements. The abnormal site is usually restricted to one or two specific vertebrae or a hip. However, multiple vertebral involvement can be difficult to distinguish from generalized causes described below.

Paget’s disease

PD commonly affects the lumbar spine and hips and has a declining UK age-adjusted prevalence of 2.5% and 1.6% for men and women, respectively [30]. PD, often asymptomatic for many years before diagnosis, explains 1.4% of incidental high BMD values [5]. Excess disorganized woven and lamellar bone expands bone size and raises density, increasing risk of deformity and fracture. PD may be monostotic (affecting an isolated vertebra) and, after the pelvis, most commonly affects lower lumbar vertebrae [31].

Tumours

Important not to miss, these most commonly occur as osteosclerotic secondary deposits from primary malignancies, e.g. prostate. Breast metastases classically cause osteolytic lesions, but can be osteosclerotic [32], as can gastric [33], colonic [34] and cervical [35] metastases. Increased BMD at an isolated vertebra can reflect a spinal osteoblastoma [36], Ewing’s sarcoma [37], carcinoid [38], haemangioma [39] or plasmacytoma [40], both of which can calcify [41, 42], and Hodgkin’s disease (5.8% of patients have spinal involvement, but osteosclerotic lesions are rarer than osteolytic) [43, 44]. Skeletal complications of radiotherapy can increase BMD, e.g. pathological fractures and secondary neoplasms. However, spinal osteoradionecrosis does not generally increase BMD, as marrow is replaced by lower density fat [45].

Tuberous sclerosis

Tuberous sclerosis is a rare, autosomal dominant disorder (OMIM 191100) of dysfunctional hamartin and tuberin production, with skeletal manifestations including bone cysts, skull and digital sclerosis and scoliosis [46]. Cortical thickening and increased bone density have been reported on plain radiographs, but DXA values have not been evaluated [47]. Learning difficulties, seizures, cardiac rhabdomyomas, haematuria from renal angiomyolipomas and dermatological features manifest variably [48].

SAPHO syndrome

SAPHO syndrome is rare, poorly understood and possibly explained by infection (Propionibacterium acnes). With features similar to the SpAs, up to half of patients suffer vertebral involvement (more frequently thoracic than lumbar) including osteosclerosis, hyperostosis, paravertebral ossification and rarely vertebral collapse [49, 50]. Case series focus on MRI and CT assessment rather than DXA, but BMD is likely to be elevated.

Generalized abnormalities causing high BMD measurements: acquired

Osteosclerosis (Greek etymology: osteo—bone, sclerosis—hardening of a tissue) generally occurs diffusely within the axial skeleton, although focal patterns may also occur secondary to exaggerated trabecular and/or cortical bone formation.

Fluorosis

Fluoride causes diffuse axial osteosclerosis with ligamentous calcification, periostitis and vertebral osteophytosis and has been associated with excessive tea and toothpaste consumption and was historically trialled as an osteoporotic therapy [51–54]. Tea leaves accumulate fluoride absorbed from the soil. Bone turnover markers [ALP, osteocalcin and C-terminal cross-linking telopeptides of type I collagen (CTX)] and BMD can be elevated [Z-scores +14 (lumbar), +7 (hip) but −0.6 (distal radius)], with enhanced cancellous bone formation on iliac crest biopsy [55]. Renal calculi have been associated [55]. Fluoride treatment does not reduce vertebral fracture risk [56, 57].

Renal osteodystrophy

Osteomalacia and soft tissue calcification are common, but renal osteodystrophy may be associated with regions of excessively mineralized bone tissue affecting the ribs, pelvis and spine. Osteosclerosis can produce the classical rugger-jersey spine X-ray appearance, characterized by sclerotic bands along multiple superior and inferior vertebral endplates with relative central lucency [58, 59].

Acromegaly

Untreated acromegaly is characterized by increased bone turnover. Excess growth hormone and insulin-like growth factor 1 (IGF-1) are anabolic, predominantly affecting cortical, rather than trabecular, bone (so increasing femoral rather than lumbar BMD) [60, 61]. However, reported hip Z-scores of +1.3 probably reflect anabolic attenuation by concurrent hypogonadism [62]. BMD changes may persist during disease remission [63].

Hepatitis C-associated osteosclerosis

Since 1992, diffuse acquired osteosclerosis, with characteristic cranial sparing, has been reported in fewer than 20 cases globally associated with HCV infection [64–79]. In addition to markedly elevated ALP, IGF proteins are apparently elevated, promoting bone formation, increasing osteoprotegerin (OPG) and reducing receptor activator of nuclear factor-κβ ligand (RANKL) levels [69, 80]. Remarkably, in one case report, lumbar spine and femoral neck T-scores of +5.5 and +15.9, respectively, fell over 7 years to +0.5 and +4.0 after successful ribavirin and IFN antiviral treatment [81]; the underlying mechanism remains unclear.

Myelofibrosis

Myelofibrosis is a rare chronic myeloproliferative disorder of bone marrow fibrosis causing marked splenomegaly and osteosclerosis, with an incidence of 0.21/100 000 person-years [82]. Small, sharp bone spicules develop within the bone marrow cavity, increasing BMD (Z-scores ranging from +2 to +6) and bone turnover in one case series of four men [83].

Mastocytosis

A disease of widespread mast cell tissue infiltration, mastocytosis has been associated with both osteoporosis and osteosclerosis. Osteosclerosis is reported in more severe disease associated with higher serum tryptase levels and higher bone turnover [84–87]. The mechanisms are poorly understood, but severe disease, with greater histamine production, may stimulate osteoblastic bone formation, while tryptase may increase OPG, reducing osteoclast activity, favouring osteosclerosis rather than osteoporosis [85, 88]. Disordered serotonin synthesis, also a feature, does not explain BMD variations [89].

Oestrogen implants

Historical use of long-term (i.e. >14 years) high-dose oestradiol implant therapy in women following surgical menopause has been associated with increased BMD in a handful of cases, with mean (s.d.) spinal and femoral neck T-scores of +1.7 (±2.0) and +1.2 (±1.4), respectively [90]. Histomorphometry suggests anabolic skeletal effects through increased osteoblastic activity.

Generalized abnormalities causing high BMD measurements: inherited

Several rare genetic disorders with skeletal effects, collectively termed sclerosing bone dysplasias and osteopetroses, are associated with generalized increased BMD [91]. Unlike spondylosis affecting multiple vertebrae, these will elevate hip as well as lumbar spine BMD. However, changes in bone structure and quantity have variable effects on fracture risk. In addition to a clinical separation based on increased or decreased fracture risk, a biological separation can be made into disorders in which (i) bone resorption is depressed (Table 2), (ii) bone formation is enhanced (Table 3) and (iii) balance is disturbed between bone formation and resorption (Table 3).

Table 2.

Osteopetrotic conditions; the gene defects, function and clinical characteristics

Condition	OMIM	Inheritance	Gene	Mutation	Protein	Function	Symptoms	Reference
Severe/neonatal/ infantile/autosomal recessive osteopetrosisa	259700, 604592	AR	TCIRG1	Loss of function	T cell, immune regulator 1, H+ transporting, lysosomal subunit A3 of V-ATPase pump	Acidification of the resorption lacuna	Fractures, infections (e.g. osteomyelitis), macrocephaly, frontal bossing, neurological symptoms, CN compression, blindness, deafness, delayed tooth eruption, haemopoietic failure, death (usually before age 10).	[92–94]
	602727	AR	CLCN7	Loss of function	Chloride channel	Acidification of the resorption lacuna
	607649	AR	OSTM1	Loss of function	Osteopetrosis-associated transmembrane protein 1	β-subunit for CLC-7
	602642	AR	RANKL/TNFSF11	Loss of function	Receptor activator for nuclear factor-κβ ligand/TNF (ligand) superfamily, member 11	Osteoclastogenesis, resorption, survival	Osteoclast poor osteopetrosis. Fractures, hydrocephalus, nystagmus, seizures, hypersplenism, less severe course than TCIRG1, CLCN7, OSTN1 mutations.	[95]
	603499	AR	RANK/TNFRSF11A	Loss of function	Receptor activator for nuclear factor-κβb	Osteoclastogenesis, resorption, survival
Intermediate autosomal recessive osteopetrosis	259710	AR	CLCN7	Partial loss of function	Chloride channel	Acidification of the resorption lacuna	Onset in childhood, fractures, short stature, cranial nerve compression.	[93, 96]
	259700, 611497	AR	PLEKHM1	Loss of function	Pleckstrin homology domain-containing family M (with RUN domain) member 1	Vesicular trafficking	Osteopetrosis of the skull only (L2–L4 T-score −2.3). Fractures. Raised osteocalcin.	[97]
Osteopetrosis with renal tubular acidosis	259730, 611492	AR	CAII	Loss of function	Carbonic anhydrase II	Intracellular acidification	Developmental delay, short stature, CN compression, blindness, dental complications, fractures, maintained haemopoietic function.	[92, 93]
Osteopetrosis with ectodermal dysplasia and immune defect (OLEDAID)	300301	XL	IKBKG	Loss of function	Inhibitor of kappa light polypeptide gene enhancer in B cells, kinase gamma (NEMO)	Unknown	Lymphoedema, severe infections, no teeth, skin abnormalities, early death.	[94]
Leucocyte adhesion deficiency syndrome (LAD-III) and osteopetrosis	612840	AR	Kindlin-3/FERMT3	Loss of function	Kindlin-3		Bacterial infections, bleeding, osteopetrosis, hepatosplenomegaly.	[98]
	612840	AR	CalDAG-GEF1	Loss of function	Calcium and diacylglycerol-regulated guanine nucleotide exchange factor 1	Cell adhesion		[99]
Late-onset osteopetrosis (Albers–Schönberg disease) ADOII	166600	AD	CLCN7	Dominant negative effect	Chloride channel	Acidification of the resorption lacuna	Classic radiographic features, fractures, nerve compression, osteomyelitis, dental complications.	[93, 100–103]
Pycnodysostosis	265800, 601105	AR	CTSK	Loss of function	Cathepsin K	Collagen degradation	Delayed cranial suture closure, short stature and phalanges, dental abnormalities, fractures.	[104–106]
Osteopoikilosis	155950	AD	LEMD3/MAN1	Loss of function	LEM domain-containing 3	LEMD3 antagonizes the BMP and TGF-β signalling pathways	Benign, incidental osteosclerotic foci (can mimic metastases)c.	[93, 107, 108]
Melorheostosis	155950	AD	LEMD3/MAN1	Loss of function	LEM domain-containing 3		Characteristic radiographic featuresd, soft tissue symptoms.
Osteopathia striatae with cranial stenosis	300373	XL	WTX	Loss of function	Wilms tumour gene on the X chromosome	Wnt signalling suppression	Macrocephaly, CN compression, cleft palate, skull/long bone sclerosis in females. Usually lethal in males.	[109]

XL: X-linked; CN: cranial nerve.

^aARO incidence is 1/200 000–300 000 live births [94]. ^bAs well as an osteoclast poor ARO phenotype, RANK mutations have also been linked to the Paget’s-like diseases (familial expansile osteolysis, expansile skeletal hyperphosphatasia and early-onset PD) [110, 111]. ^cWhen associated with connective tissue naevi, dermatofibrosis lenticularis disseminata then termed Buschke–Ollendorff syndrome [93, 107, 112]. ^dAsymmetric ‘flowing hyperostosis’ or ‘dripping candle wax’. Approximately 200 cases described to date. Soft tissue changes (hypertrichosis, fibromas, haemangiomas and pain) associated with radiographic features in sclerotome. Contractures can develop [93, 107, 108, 113]. ^eCan occur in combination with focal dermal hypoplasia, skin pigmentation, hypoplastic teeth, syndactyly, ocular defects and fat herniation through skin and is known as Goltz syndrome [109, 114–116].

Table 3.

Inherited HBM conditions due to enhanced bone formation, or disturbed formation and resorption; the gene defects, function and clinical characteristics

Condition	OMIM	Inheritance	Gene	Mutation	Protein	Function	Symptoms	Reference
Increased bone formation
Sclerosteosis	269500	AR	SOST	Loss of function	Sclerostin	Osteoblast Wnt signalling inhibitor	Cutaneous digital syndactyly, excessive height. Skull/mandible thickening, toria, CN palsies (incl. neonatal). Headaches, raised ICP, coning. Back/bone pain. Fracture resistance.	[117, 118–120]
van Buchem’s  diseaseb	239100	AR	SOSTc	Reduced function	Sclerostin	Osteoblast Wnt signalling inhibitor	No syndactyly, no excess height. Skull/mandible thickening, toria, CN palsies. Headaches, back/bone pain. Fracture resistance.	[117, 121, 122]
LRP5 HBM	603506	AD	LRP5	Gain of function	LRP5	Osteoblast cell membrane co-receptor regulating Wnt signalling	Asymptomatic or toria, skull/mandible thickening, CN palsies, neuropathy, neuralgia, headaches, back/bone pain, spinal stenosis, reduced buoyancy, craniosyntosis. Fracture resistance.	[123, 124–140]
LRP4 HBM	604270	AD and AR	LRP4	Loss of function	LRP4	Impaired sclerostin–LRP4 interaction	Syndactyly, dysplastic nails, gait disturbance, facial nerve palsy, deafness.	[141]
Craniometaphyseal  dysplasia	123000	AD and AR	ANKH	Loss of function	ANK	Osteoclast-reactive vacuolar proton pump	Macrocephaly, cranial hyperostosis, CN palsies, wide nasal bridge, dental overcrowding, metaphyseal flaring.	[142, 143]
	218400
Disturbed balance between bone formation and resorption
Camurati–  Engelmann  diseased	131300	AD	TGF-β1	Probable gain of function	TGF-β	Stimulates both osteoblast and osteoclast activity	Onset before 30 years, variable phenotype. Thickened diaphyseal cortices, limb pain, fatigability, muscle weakness, waddling gait. Variably raised ALP, reduced calcium and anaemia.	[144–149]
Ghosal haematodiaphyseal syndrome	274180	AR	TBXAS1	Loss of function	Thromb-oxane synthase	Modulates RANKL and OPG expression	Impaired platelet aggregation (steroid-sensitive), anaemia. Similar to Camurati–Engelmann but metaphyses affected.	[150, 151]

OMIM®: Online Mendelian Inheritance in Man; CN: cranial nerve; ICP: intracranial pressure.

^aTori: oral exostoses, which include torus palatinus and mandibularis found in approximately 25% of a general Caucasian population [152]. ^bInitially known as hyperostosis corticalis generalisata familiaris [121, 122]. ^cA 52-kb intronic deletion downstream of SOST. ^dAlso known as progressive diaphyseal dysplasia.

Decreased bone resorption

Osteopetroses (Greek etymology: petro—to turn to stone) are rare genetic conditions of reduced osteoclastic bone resorption. Defective bone remodelling during growth induces skeletal sclerosis and abnormally dense but brittle bones, first described by Albers-Schönberg as marble bone disease [92, 153]. Osteopetrosis is classified by clinical severity (Table 2); autosomal dominant osteopetrosis (ADO) was historically subdivided into type I and type II. ADOI, subsequently identified as secondary to an LRP5 (low-density lipoprotein receptor-related protein 5) mutation [123] (discussed later), is not a primary osteoclast disease, is not characterized by bone fragility and is better not considered as an osteopetrosis. Two osteopetroses pertinent to adulthood are discussed below.

Autosomal dominant osteopetrosis II

ADOII (Albers–Schönberg disease) is caused by a CLCN7 mutation with penetrance between 60% and 80%, giving a varied clinical phenotype, including detection as an incidental radiographic finding [154]. Prevalence is estimated at between 0.2 and 5.5/100 000 [155, 156]. The phenotype can include facial nerve palsy, visual loss (in 5–25%), carpal tunnel syndrome, hip OA (in 7%), increased fracture risk and delayed fracture healing, osteomyelitis (in 10–13%), particularly of the mandible, dental abscesses (10%) and deep decay (36%) and in extreme cases bone marrow failure (∼3%) [93, 100–103]. In one case series of 94 CLCN7 mutation cases, almost every adult (98%) had experienced a fracture (including half of their hip), with a third having fractured more than once (five had >15 fractures) [102]. Among another 42 cases from 10 families, age range 7–70 years, the mean number of fractures per person was 4.4 [103]. However, these case series were not performed systematically so patterns are difficult to generalize.

Radiographs feature (i) vertebral end-plate thickening (rugger-jersey spine), (ii) bone-within-bone, particularly in the pelvis, and (iii) transverse sclerotic bands within the distal femorae [100, 103]. However, the radiological phenotype is not ubiquitous (∼60–90%) [155, 157]. DXA BMD Z-score ranges from +3 to +15 [100, 102]. The CLCN7 protein functions as a voltage-gated Cl⁻/H⁺ ion channel and is found in lysosomes and on the ruffled border of osteoclasts. By acid efflux, it facilitates inorganic bone matrix dissolution [158]. Multiple mutations have been identified in association with the range of osteopetrotic phenotypes [159–161].

Pycnodysostosis

First described in 1962 and said to be the malady of both Toulouse-Lautrec and Aesop (of fable renown) [162–164], pycnodysostosis is caused by defective enzymatic degradation of organic bone matrix due to an autosomal recessive mutation in the gene coding cathepsin K [104]. To date, 27 mutations have been reported among fewer than 200 cases globally [104–106]. Secreted by osteoclasts, cathepsin K cleaves type I collagen [165]. The characteristic bone dysplasia includes skull deformities, underdeveloped facial bones with micrognathia, beaked nose, short stature and phalanges, dental caries, persistence of deciduous teeth and abnormally dense but brittle bones [93, 104–106, 166]. Interestingly, understanding of pycnodysostosis has prompted development of a novel class of antiresorptive therapy currently in trial (e.g. odanacatib) [167] (Table 4).

Table 4.

Examples of how understanding HBM conditions has helped inform development of new osteoporosis therapies

HBM condition	Molecular target	Drugs in development	Reference
Pycnodysostosis	Cathepsin K	Cathepsin K inhibitors:
		Odanacatib (Phase III trial)	[167]
		Balicatib (trials discontinued due to dermatological side effects)	[168, 169]
Sclerosteosis and van Buchem’s disease	Sclerostin	Anti-SOST antibodies	[170, 171]
LRP5 HBM and osteoporosis pseudoglioma syndrome (OPPG)	Inhibition of natural antagonists of osteoblastic Wnt signalling	Glycogen synthase kinase-3β (GSK3β) inhibitors	[172]
		Dickkopf 1 (Dkk1) antibodies	[173, 174]
		Secreted frizzled-related protein-1 (Sfrp1) inhibitors	[175]

Increased bone formation

Sclerosteosis and van Buchem’s disease

Sclerosteosis and van Buchem’s disease are clinically similar conditions of generalized enhanced bone formation, increased bone strength and resistance to fracture due to reduced levels of sclerostin [117]. It is thought that mechanical loading reduces osteocytic production of sclerostin, permitting activation of osteoblastic Wnt signalling and bone formation [176]. At least three pharmaceutical companies are currently developing anti-sclerostin antibodies [170, 171] (Table 4). Loss-of-function SOST gene mutations cause sclerosteosis, whereas a 52-kb intronic deletion downstream of SOST, thought to disrupt post-transcriptional sclerostin processing, results in the milder phenotype of van Buchem’s disease. Sclerosteosis causes gigantism, mandible enlargement, torus palatinus and mandibularis, which complicate tooth extractions [118, 177]. Calvarial overgrowth compresses cranial nerves, particularly facial nerves, sometimes from infancy; in one series, 83% of 63 adults had recurrent facial nerve palsies [118]. Hearing loss and headaches are common; craniotomy to alleviate raised intracranial pressure and sudden death by coning is not uncommon [118, 178]. Cutaneous syndactyly of fingers (present in 76%) and toes is an important defining feature, often accompanying dysplastic or absent nails and camptodactyly [118, 178, 179]. Sclerosteosis is progressive, which may cause bone and back pain requiring spinal decompression [118].

van Buchem’s disease is milder than sclerosteosis, importantly without syndactyly or gigantism [117, 178]. Cranial nerve impingements and hearing loss remain common [180]. Management is generally limited to surgical bone removal, however, glucocorticoids have been used to reduce high bone turnover in an isolated case report [181].

LRP5 high bone mass

Ten activating LRP5 mutations affecting 23 families globally have now been reported [123, 124–139]. Initially cases were described as asymptomatic, with mandible enlargement, osseous tori, a marked resistance to fracture (e.g. in car accidents), thickened cortices on radiographs (without reduced haemopoietic capacity), normal biochemistry and BMD Z-scores of +3 to +8 [124, 182]. However, subsequent case reports describe complications secondary to bone overgrowth: nerve compression causing deafness, cranial nerve palsies, congenital strabismus, sensorimotor neuropathy, spinal stenosis, paresthesias and trigeminal neuralgia [127, 128], in addition to headaches, bone pain and reduced buoyancy [126, 127]. The G640A mutation is the only one to link LRP5 with craniosynostosis requiring craniotomy, developmental delay and a profoundly dysmorphic and pathological phenotype including ventricular septal defect (VSD) [129]. Osteocalcin levels are raised or normal [126, 127, 182]. LRP5 codes for an essential cell membrane co-receptor within the Wnt signalling pathway, regulating osteoblastic bone formation [140]. Conversely, inactivating LRP5 mutations causes autosomal recessive osteoporosis pseudoglioma syndrome (OPPG) [183].

Unexplained high bone mass

There remains a population, even after exclusion of all of these listed conditions, with a sporadic finding of generalized raised BMD (Z-score ≥+3.2 at either L1 or hip) on routine DXA scanning with unexplained high bone mass (HBM) in whom fracture risk is not increased, associated with clinical characteristics suggestive of a mild skeletal dysplasia, namely poor buoyancy, mandible enlargement, extra bone at the site of tendon and ligament insertions, broad skeletal frame and larger shoe size, as well as increased BMI [5]. Considered to be relatively benign, this picture explains 35% of incidental findings of raised BMD on routine DXA scanning. As 41% have a first-degree relative with a similar phenotype, it is thought to be an inherited condition. Research is currently under way to identify the genetic cause and fully evaluate the associated phenotype, e.g. metabolic, muscular and joint characteristics, to inform clinical management.

Recent findings suggest that HBM is characterized by increased trabecular BMD and by alterations in cortical bone density and structure, leading to substantial increments in predicted cortical bone strength. Neither trabecular nor cortical BMD appear to decline with age in the tibia of HBM cases, suggesting that attenuation of age-related bone loss in weight-bearing limbs may contribute to their bone phenotype [184]. Furthermore, body composition assessment suggests that HBM is associated with a marked increase in fat mass, particularly android fat, in women but not men [185]. Although elevated BMI is not a recognized feature of skeletal dysplasia, interestingly, a similar finding has been reported in families of HBM due to an activating LRP5 mutation [186].

Finally, studying HBM may improve our understanding of OA. An inverse relationship between osteoporosis and OA is well documented, with higher hip and/or lumbar spine BMD in individuals with radiographic OA [187–190]. However, osteophytes can artefactually increase measured BMD [9] and, counterintuitively, fracture risk is not reduced in OA [191, 192]. Potential mechanisms linking increased BMD with OA include (i) increased subchondral bone stiffness increasing articular cartilage stresses and damage [193], (ii) activation of the Wnt signalling pathway, thought to have a role in both joint formation and maintenance of joint homeostasis in later life [194] (supported by β-catenin upregulation in knee joint cartilage prior to joint replacement [195]) and (iii) molecular cross-talk between bone and cartilage arising through increased permeability of the bone–cartilage interface [196, 197]. Large joint OA has been reported in ADOII and LRP5 HBM [103, 127, 138], and unexplained HBM has recently been associated with an increased prevalence of joint replacement [198], suggesting that increased OA risk may represent a further, hitherto unrecognized, consequence of elevated BMD.

Investigation and management of a raised BMD

Initial inspection should classify BMD increases as focal or generalized (spine, hip or both). Focal increases in BMD should be carefully inspected for osteoarthritic changes, which if clearly visible require no further imaging. Otherwise, anteroposterior (AP)/lateral lumbar spine ± pelvis plain X-rays are initially recommended with routine bone biochemistry and inflammatory markers. MRI may be required, particularly if examination prompts doubt regarding spinal cord compression or X-rays raise the possibility of malignancy. Lateral DXA can help with vertebral fracture assessment. Suspected malignancy may require mammography, isotope bone scan, prostate assessment and tumour markers. ALP is usually lowered in hypophosphatasia and raised in active PD, although up to 5% will have a normal ALP in PD [199].

Generalized increased BMD affecting both spine and hip are less commonly seen and the differential diagnosis is wide. Outpatient clinic assessment should include questioning regarding fluoride exposure, hepatitis C risk factors, headaches, bone pain and in women historical oestrogen implant use, plus examination for stigmata of acromegaly, bone overgrowth, nerve compression, splenomegaly (in haemopoietic failure) and dysmorphism suggestive of a mild skeletal dysplasia associated with unexplained HBM. A careful fracture history is essential, including the family history. Blood tests should include bone biochemistry, renal function, full blood count (FBC) and clotting studies, liver function and hepatitis C serology, plus potentially serum fluoride levels, IGF-1 ± an oral glucose tolerance test if acromegaly is suspected and serum tryptase if mastocytosis is suspected. Bone turnover markers (P1NP and serum CTX) may be useful.

Potentially relevant plain radiographs include AP/lateral lumbar spine, pelvis, bilateral femorae and lateral skull. In ADOII, radiographs show the classic rugger-jersey spine due to vertebral end-plate thickening, bone-within-bone often seen in the pelvis and transverse sclerotic bands within the distal femur [100, 103]. DXA examination showing low distal radius BMD would support the diagnosis of fluorosis [cranial sparing on whole-body DXA scanning, if available, would support hepatitis C-associated osteosclerosis (HCAO)]. Hip and lumbar spine DXA scans in first-degree relatives will help identify relatively asymptomatic inherited HBM conditions. If specific characteristic features suggest a monogenic disorder such as osteopetrosis or sclerosteosis, referral to local clinical genetic services for counselling and genotyping should be considered depending on the severity of symptoms and the family history.

If an inherited condition of increased bone formation is suspected, a number of investigations may be helpful in establishing the severity of the phenotype. Visual field assessment and formal audiology are important as cranial nerve impingement can be managed by surgical decompression. For similar reasons, CT/MRI skull, MRI spine and nerve conduction studies may be helpful. Assessment by dental and/or maxillofacial specialists may be needed. Examination should include cardiovascular examination, and if a severe LRP5 mutation is suspected, cardiac echocardiography may be needed to exclude VSD. An approach to investigating high BMD measurements is summarized in Fig. 2.

Fig. 2.

Flow diagram to guide the investigation and management of raised BMD identified on DXA scanning.

Ca²⁺: Calcium; PSA: prostate specific antigen; PV: plasma viscosity; U+E: urea and electrolytes; PO₄: phosphate; OGTT: oral glucose tolerance test; LFTs: liver function tests; P1NP: N-terminal propeptides of type I procollagen. *Up to 5% with PD will have a normal ALP [161]. Potential diagnoses are each given a superscript digit, to which the investigations then relate.

Conclusion

A BMD T/Z-score >+2.5 does not generally indicate normal bone density, but is usually caused by an artefactual increase in BMD secondary to lumbar spondylosis which is readily identifiable from inspection of the DXA scan image. However, high BMD measurements may arise from a genuine increase in bone mass. This may be caused by a focal abnormality within the DXA field, such as a Pagetic lumbar vertebra, or a generalized skeletal abnormality resulting from acquired osteosclerosis, or rarely a genetic mutation leading to a sclerosing bone dysplasia. The most common form of sclerosing dysplasia is the currently unexplained HBM phenotype, characterized by a mild skeletal dysplasia; unlike the osteopetroses, this does not convey an increase in fracture risk.

Knowledge of rare genetic skeletal dysplasias has helped guide innovative treatments for osteoporosis (Table 4), e.g. from our understanding of pycnodysostosis, odanacatib was developed [167], as were anti-sclerostin antibodies from our experience of sclerosteosis and van Buchem’s disease [170]. Yet, much HBM remains unexplained, better appreciation of which may translate into improved understanding of bone regulation and new therapeutic targets for future osteoporosis therapies, as well as aiding management through greater understanding of associated comorbidities.

Here we have presented a classification for the potential causes of a raised BMD detected by DXA scanning as part of normal clinical practice. This classification should help guide clinical evaluation and diagnosis when the DXA scan is interpreted within the context of the clinical history.