If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Medicine, Faculty of Medical & Health Sciences, University of Auckland, Private Bag 92019, Auckland Mail Centre, Auckland 1142, Aotearoa, New Zealand.
Paget's disease is a chronic focal high turnover bone disorder that is primarily present in middle-aged or older adults. It seems to be restricted to humans and has no clear parallels with other diseases. Although much has been learnt about its pathology and epidemiology, and treatment is now highly effective we still lack a complete understanding of its etiology and biology. This review focusses on the natural history of the disorder, in particular its changing epidemiology, recent discoveries about its genetic basis and current approaches to diagnosis and treatment. While there is strong evidence for genetic predisposition to Paget's disease, there is also compelling evidence that it is becoming less prevalent, the age of patients at presentation is increasing and that the extent of skeletal involvement is diminishing, implying that there is an important, but as yet unidentified, environmental factor in its etiology. Contemporary patients are typically elderly and have few bones involved. Treatment with potent intravenous bisphosphonates provides prolonged remission and many will require only once in a lifetime treatment.
In 1876 the London physician, Sir James Paget, published his famous monograph concerning a chronic condition of bone that he named osteitis deformans. Paget's keen observations, made over several years, accurately described the distinctive clinical features of the disorder: bone pain, enlarging skull size, deafness, fractures, deformity and the late development of malignancy [
]. Since that time, thanks to technologies that were not available to Paget, we understand better its pathophysiology and we now have effective treatments, but there are still many unsolved mysteries about the nature of the disease, its etiology and its natural history.
1. Clinical Features and Natural History
Paget's disease is a chronic focal bone disorder primarily presenting in middle-aged or older adults (though, as discussed below, the presentation is often earlier in those with an identifiable genetic mutation). The disease appears to arise more or less simultaneously in one or more skeletal sites and remains restricted to these sites. Renier & Audran [
] reported new skeletal lesions developing in 10 patients a mean 23 years after their initial lesions were identified. However, in the majority of these cases the new lesions were recognised by skeletal scintigraphy, a more sensitive technique that was not used at the initial identification, so one cannot be sure that these lesions were genuinely ‘new’. In a large study of 100 patients who had repeat skeletal scintigraphy a mean 5½ years after diagnosis, none had developed new lesions [
]. Although the ‘birth’ of pagetic lesions is not a phenomenon that has been observed in close detail, there is little convincing evidence that lesions appears asynchronously. The one exception to this rule is that the disease can be transmitted to new skeletal sites by the accidental use of pagetic bone for surgical bone grafting [
In long bones the disease first appears in the region of the proximal epiphysis and advances along the shaft at a rate of ~8 mm/yr The leading edge of this advance is often visible as a V-shaped ‘lytic wedge’ reflecting osteoclastic resorption (Fig. 1A ). The same phenomenon in the skull gives rise to the radiographic appearance of osteoporosis circumscripta. Having reached its maximal extent and activity, the disease then appears to remain more or less stable for many years with relatively small fluctuations in bone turnover markers. Less well documented, and unexplained, is the observation in some patients of a late phase in which the turnover of pagetic bone declines spontaneously, leaving sclerotic but metabolically inactive bone.
Fig. 1A A ‘lytic wedge’ of osteoclast-led resorption in the tibia. Note the thin cortex which makes this phase vulnerable to fracture.
B Paget's disease affecting the right femur and right hemipelvis causing secondary osteoarthritis at the hip joint and acetabular protrusion.
C Monostotic Paget's disease affecting the third lumbar vertebra. Note that the bone is enlarged.
D An 85 year old woman with PDB affecting the skull and left fibula (arrowed) before (left) and after a 6 month course of oral alendronate (40 mg/day). Note the marked reduction in isotope uptake.
In a high proportion of people PDB is discovered incidentally on biochemical testing or radiography, and although many of these probably do have pain at least 20–25% are genuinely asymptomatic [
]. Pain is the commonest symptom: pagetic bone pain is typically worse at rest (often at night or early after rising) and is relieved by movement whereas pain from osteoarthritis secondary to juxta-articular Paget's disease (Fig. 1B) tends to be worse with movement. In the elderly, in whom musculoskeletal symptoms are very common, it can be difficult to be certain of the exact contribution of PDB to pain. Other symptoms caused by complications include pathological fracture (most commonly in lytic disease in long bones), painful fissure fractures, bone enlargement and deformity. PDB in the skull can cause deafness (usually from involvement of the otic capsule) and increasing hat size. Very rarely vertebral involvement can cause an acute myelopathy (most commonly due to a vascular ‘steal’ rather than cord compression).
2. Investigations
At the time of diagnosis, radiographs typically show a mixed picture of lysis and sclerosis that give Paget's disease its distinctive radiographic appearance. The increased rate of remodelling is reflected by increases in bone turnover markers. All bone formation and resorption markers are affected to similar degree, with the exception of plasma osteocalcin, which is low relative to other formation markers such as alkaline phosphatase (ALP) or the procollagen peptides.
In the untreated state, the degree to which bone turnover markers are elevated is dependent on both the activity and the extent of the disease. Paget's disease is often first detected by the finding of an elevated ALP level with normal activities of other liver-derived enzymes. The secular trend toward less extensive disease (discussed below) means that this has become a less reliable screen - as ALP may not be elevated above the normal range in patients with disease of limited extent. Of the newer bone turnover markers, procollagen-1 N-propeptide seems to perform best, in terms of being elevated above normal even if the disease is of limited extent, response to bisphosphonate treatment and early detection of relapse [
Bone scintigraphy is the best method of determining the extent of skeletal involvement. This can be quantitated either as the total number of bones affected or the proportion of the skeleton involved, estimated using methods such as that of Coutris et al. [
]. The radiographic and scintigraphic appearances are characteristic (Table 1, Fig. 1C and D). Any bone in the skeleton can be involved, but some more commonly than others (pelvis, spine, sacrum, skull, femora and tibiae).
Table 1Radiographic features of Paget's disease.
•
Classical triad
-
Thickening of the cortex
-
Accentuation of the trabecular pattern
-
Increased size of bone
•
Cyst-like areas
•
Skull involvement
-
Inner and outer table involved
-
Diploic widening
-
Osteoporosis circumscripta - well-defined lytic lesion in skull
-
“Cotton wool” appearance of thickened calvarium
-
Basilar invagination with encroachment on foramen magnum
-
Sclerosis of skull base
•
Long bones
-
V-shaped defect of advancing lytic tip in diaphysis
-
Lateral curvature of femur
-
Anterior curvature of tibia
-
Fissure fractures
•
Pelvis
-
Thickened trabeculae in sacrum, ilium
-
Rarefaction in central portion of ilium (looks like a large lytic lesion)
-
Thickening of iliopectineal line
-
Acetabular protrusion with secondary degenerative joint disease
•
Spine
-
Coarse trabeculations at periphery of bone
-
“Picture-frame vertebra”
-
Enlarged vertebral body; reinforced peripheral trabeculae/lucent center
The most important differential diagnoses are fibrous dysplasia, chronic osteomyelitis and metastases. Biopsy is rarely needed to establish the diagnosis.
3. Pathology
Paget's disease is characterised by highly exaggerated remodelling in affected bones, with abnormalities in all phases of the remodelling cycle. The primary abnormality is believed to lie in the regulation of osteoclasts (the cells that resorb and remove bone), which are increased in both number and size and are also hypernucleated (up to 50 nuclei per cell). In response to the rapid bone resorption, bone formation becomes greatly accelerated: tetracycline-based histomorphometric estimates suggest that it is 6–7 fold greater than normal [
]. The new bone formation is chaotic, with loss of its normal lamellar pattern, so that despite affected bones being larger and more sclerotic than normal, the bone is poor quality – hence the tendency to deformity and fracture.
In vitro studies of bone marrow cells from affected bone have shown that pagetic osteoclast precursors have an abnormal ‘phenotype’ in that they are hypersensitive to several osteoclastogenic factors including RANKL, TNFα and 1,25(OH)2D3: they will form multicellular osteoclasts at concentrations of these factors 10 to 100-fold lower than required for the formation of normal osteoclasts [
]. The level of TAFII-17, a component of the TAFIID transcription complex that binds the vitamin D receptor, is increased in pagetic osteoclast precursors and may be responsible for the hypersensitivity to 1,25(OH)2D3 [
]. Other abnormalities have been noted in serum and peripheral blood cells, including increased levels of interleukin 6 (IL-6) and interferon γ, but these may be epiphenomena of high bone turnover [
The study of exhumed skeletons has suggested that the earliest cases of Paget's disease were seen in Western Europe (predominantly Britain) in the Roman period, around 1–400 CE [
]. By the second half of the 19th century – an era of rapid industrialization and urbanization – florid cases with extensive skeletal involvement, such as those described by Paget, may have been reasonably plentiful. Barker & Gardner [
] documented that a high proportion of the death certificates of English and Welsh people born in the 1850s mentioned ‘Paget's disease’, but there was a progressive decline in cohorts born through the 1920s (Fig. 2). There was a similar secular decline in the number of deaths due to adult-onset osteosarcoma (presumed mainly to arise in pagetic bone). Nonetheless, the disease remained prevalent in Europe: Schmorl in his 1932 autopsy study from Dresden [
] found Paget's disease in 2–3% of subjects aged 50–60 years and 4–8% of those aged 80–90 years (the higher values in these ranges being men, and the lower women).
Fig. 2Proportion of death certificates issued in England & Wales that mentioned Paget's disease in successive cohorts born between 1880 and 01 and 1890–01. Note that at an age of death of 72½ (dotted line) the proportion had fallen from 25% to 9%. Data adapted from Gardner & Barker
Since the 1970s the radiographic survey has become the standard way of assessing the prevalence of Paget's disease. In such surveys a large number of sequential radiographs of the pelvis, taken for any reason, are examined for signs of Paget's disease. In ~70% of affected people the disease involves one or more of the lumbar spine, sacrum, pelvis or proximal femur, so can be seen on an abdominal radiograph. Such surveys were undertaken in the United Kingdom, continental Europe, North America, Australia, New Zealand and South Africa. They showed not only marked differences in prevalence between countries, but also regional differences within countries. The United Kingdom had the highest country prevalence, and it was particularly high in the north-west of England (‘the Lancashire focus’). In the USA, the prevalence was highest in the north-eastern states [
It is commonly stated that PDB is rare in populations of non-European descent, but few formal prevalence surveys have been undertaken outside North America, Australasia or Europe. In South Africa the prevalence of PDB in people of black African descent was similar to that of people of European descent [
] has been generally interpreted as evidence of genetic predisposition. However, in recent times in New Zealand our group has reported that a substantial proportion of newly diagnosed patients are of Asian descent [
In several countries repeat radiographic surveys were undertaken about twenty years after the first, and in almost all previous high prevalence areas there has been a substantial fall in the proportion of the population affected [
]. Over a similar period there has also been a remarkable reduction in the severity of the disease, as assessed by the proportion of the skeleton involved [
] and severe polyostotic disease, of the type recognised by Paget, is becoming something of a rarity. This is not because patients are being diagnosed earlier in life and having medical intervention – the average age at diagnosis has been steadily increasing [
Much recent research has focused on identifying genetic causations or predisposition to Paget's disease, but the as yet unexplained reduction in prevalence and severity of the disease points to there being a strong environmental component to its etiology.
5. Genetics of Paget's Disease
The familial occurrence of the disease was first recognised around 1926 and recent estimates suggest that 15–30% of individuals with PDB have one or more affected relatives. In the past two decades there has been intense interest in exploring the genetic contribution its etiology.
5.1 SQSTM1 and p62
The most significant discovery has been that ~30% of familial PDB is associated with dominantly-inherited mutations in the SQSTM1 gene [
]. Compared to affected individuals without mutations, those with SQSTM1 mutations tend to have a more severe phenotype - presenting at an earlier age and having more extensive disease [
] for a current list), but one, p.P392L, is by far the most common, and has been found in populations around the world. The phenotype of the few people so far described who have bi-allelic mutations does not seem to be any more severe than those with a single mutation. Around 5% of those with sporadic (non-familial) PDB also carry SQSTM1 mutations.
Early studies suggested a high penetrance, with ~80% of family members inheriting a mutation developing the disease by the age of 70 [
]. However, SQSTM1 was identified originally by positional cloning, for which families with a high disease penetrance were particularly informative, so these observations may not be generalisable, and moreover the penetrance may be changing. Our group studied how PDB develops in family members inheriting SQSTM1 mutations from affected parents, and found that in recent generations the phenotype is attenuated: fewer are developing the disease (or its development is delayed by 10 years or more) and the extent of skeletal involvement is also less [
] (Fig. 3). This is very much in accord with the secular trends discussed above. Even within families known to carry SQSTM1 mutations there may be individuals with PDB that do not carry the family mutation, implicating other causal factors [
Fig. 3Bone scintiscans from a father and son who both carry a truncating mutation in SQSTM1. The father had severe symptomatic polyostotic disease when he was diagnosed at the age of 39 - his ALP was 1200 u/L. His son was asymptomatic when diagnosed with monostotic PDB of the sacrum at the age of 44 - his ALP was 81 u/L (normal <120 u/L).
SQSTM1 encodes the ubiquitous multidomain intracellular scaffold protein, p62, that is involved in various signalling pathways. How changes in p62 predispose to PDB is not understood fully. The protein has a number of functions, but most attention has been focussed on those that affect osteoclast differentiation, activity or survival. These include transduction of the NFΚB and the oxidative-stress induced Keap1/Nrf2 pathways, targeting proteins for degradation and the formation of the autophagosome, and apoptosis. p62 is involved in many RANKL-activated osteoclast signalling pathways. The interaction between receptor activator of NFκB ligand (RANKL) and its receptor (RANK) in osteoclasts results in signalling cascades that activate transcription factors, such as NFκB and NFATc1, that in turn promote and regulate osteoclast differentiation, activity and survival [
Autophagy is a critical pathway for the removal of damaged and aggregation-prone proteins and organelles. p62 is a cargo receptor for ubiquitin-mediated autophagy [
This is a new term for a group of conditions encompassing fronto-temporal dementia and/or inclusion body myopathy or amyotrophic lateral sclerosis/motor neurone disease, and PDB Affected tissues have in common ubiquitin-positive inclusions that contain RNA-binding proteins such as TDP-43, hnRNPA1 and hnRNPA2B1, but may also contain proteins that mediate ubiquitin-dependent autophagy such as p62, VCP and optineurin [
]. To date mutations in five genes have been associated with multisystem proteinopathy: VCP, HNRNPA2B1, HNRNPA1, SQSTM1 and MATR3. Paget's disease has been associated with all these except the last. SQSTM1 mutations associated with multisystem proteinopathy are found throughout the gene, and not clustered in the region coding for the UBA domain, although there is considerable overlap [
) and 37 patients with SQSTM1 mutations but no multisystem proteinopathy (data from Auckland, New Zealand). Paget's disease was diagnosed about 10 years earlier in VCP mutation carriers.
The genes implicated in multisystem proteinopathy encode either RNA-binding proteins, or proteins that mediate ubiquitin-dependent autophagy. The result appears to be excess assembly of RNA granules (that normally control post-transcriptional mRNA metabolism) coupled with failure to degrade them by autophagy. One small study has reported that particular alleles in certain autophagy-related genes are also associated with the risk of PDB [
Malignant transformation of Paget's disease is a feared but fortunately rare complication. In most cases the tumor is an osteosarcoma (osteoblast-derived), but there is an interesting subgroup in whom giant cell tumors develop. This syndrome, which is strongly familial, is linked to mutations in the gene ZNF687. This encodes a C2H2 zinc finger protein that is part of a transcriptional regulator complex, that is expressed in most tissues, including bone. PDB in these patients is typically polyostotic and of early onset. The tumors are commonly multifocal and on average appear 10 years after the diagnosis of PDB, though this interval can vary from 3 to 30 years [
], and there is much current interest in identifying how these polymorphisms might predispose to PDB. Several of the genes identified CSF1, TNFRSF11A, TM7SF4 and RIN3 encode proteins known to be important to osteoclast function. OPTN encodes a protein (optineurin) that is involved in the autophagy pathway, and specific mutations in OPTN have been associated with amyotrophic lateral sclerosis/motor neurone disease and dementia (but not, to date, PDB). The roles of two other genes identified by these studies are currently unknown. The protein encoded by NUP205 is involved in the transport of macromolecules between the cytoplasm and nucleus and PML encodes a nuclear tumor suppressor protein that regulates cell proliferation.
The phenotype associated with SQSTM1, VCP, HNRNPA2B1 and ZNF687 mutations is notably more severe than sporadic PDB, with an earlier age of onset and more extensive disease. In contrast, the effect of the GWAS-identified susceptibility alleles appears small. Albagha et al. genotyped PDB patients and examined the phenotype according to how many susceptibility alleles each patient had. In those without SQSTM1 mutations, the mean number of bones involved was 1.7 for those in the lowest and 1.9 in the highest tertile of risk, and there was no difference in mean age at diagnosis [
There are no naturally occurring animal models of PDB, but two groups have reported on the phenotype of genetically engineered mice with a P394L knock-in mutation (equivalent to the human P392L SQSTM1 mutation). Neither completely recapitulates the human phenotype. Daroszewska et al. [
] found focal high turnover bone lesions, predominantly affecting the lower limbs, that developed with increasing age. The phenotype was more marked in mice homozygous for the mutation than in heterozygous mice (thus differing from human PDB). The osteoclasts within lesions were larger and more nucleated than normal and some contained nuclear inclusions, similar to those observed in human PDB. Osteoclast precursors from P394L mutant mice had increased sensitivity to RANKL in vitro and generated osteoclasts that were larger and more nucleated than those from wild-type littermates. In contrast, Kurihara et al. [
] found that mice expressing p62(P394L) formed normal osteoclasts. However, mice expressing measles virus nucleocapsid protein (MVNP) in osteoclasts, with or without p62(P394L), developed pagetic-like osteoclasts and expressed high IL-6 levels. Mice co-expressing MVNP and p62(P394L) developed paget-like bone lesions in the vertebrae, but the limbs of these mice were not examined. These authors argued that the SQSTM1 mutation alone was insufficient to create a paget-like phenotype.
7. Treatment
There have been remarkable developments in treatment, based on the availability of agents that can suppress the accelerated bone turnover. The first effective therapy for this previously intractable and often disabling disorder was calcitonin, introduced in 1968 [
]. Calcitonin's effects were however incomplete and short-lived, and it was rapidly superseded by bisphosphonates. Etidronate was the first used in 1973 [
], but it was limited by toxicity (the induction of low turnover osteomalacia) when it was given in high doses for prolonged periods. Over the following 30 years, bisphosphonates of progressively greater potency and duration of action were developed and introduced into clinical practice.
There is good evidence that suppressing bone turnover with bisphosphonate treatment improves pagetic bone pain, heals lytic lesions, reduces bone turnover and restores more normal bone structure [
]. Because the long-term complications such as deformity, secondary osteoarthritis and deafness may take decades to evolve there is no proof that bisphosphonate treatment can prevent their development, although this remains plausible [
In recent years aminobisphosphonates of progressively greater potency have been introduced into clinical practice. Many different bisphosphonates, given orally or intravenously and in a variety of regimens have been used to treat PDB. All suppress bone turnover, but relapse (as indicated by rising bone turnover markers) is inevitable upon cessation of treatment. The duration of response depends in part on the extent of disease, the potency of the bisphosphonate, and the dose used. The response with widely used agents such as pamidronate, which is of intermediate potency, is often of relatively short duration.
Zoledronate is the most potent bisphosphonate tested to date. In a large randomized clinical trial, a single intravenous dose of zoledronate, induced a ‘therapeutic response’ (defined as normalization of the ALP level, or a reduction of at least 75% in the ALP excess) in 98% of patients within 6 months [
]. In a follow up study, relapse, defined as a rise in ALP to within 20% of its pre-treatment value, occurred in <1% of subjects up to 6½ years after a single intravenous infusion, though some loss of therapeutic response occurred in 12.5% [
The subjects in this clinical trial were somewhat atypical of patients with PDB now seen in clinical practice, as almost all had ALP values at least twice the upper limit of normal, indicating moderate or severe disease. In recent decades the marked secular change in the clinical presentation of PDB discussed above means that, compared to 40 years ago, patients are significantly older at diagnosis and have less severe disease (fewer bones involved with PDB). It is now common for patients with disease of limited extent to have ALP values that are not elevated above the normal range. The combination of milder disease occurring in older patients on the one hand, coupled with the use of more potent bisphosphonates with a long duration of effect on the other, raises the prospect of many patients needing only once in a lifetime treatment.
In a recent study we reported the follow-up for 6 to 10 years of a cohort of PDB patients who at a mean age of 76 years had been treated with a single intravenous zoledronate infusion. Up to 10 years after a single infusion of zoledronate, the cumulative risk of relapse (as measured by the sensitive bone turnover marker procollagen-1 N-propeptide rising above the normal range) was only 14%, but the cumulative risk of death was >50%. The implication is that a single dose of a potent bisphosphonate will be adequate for lifelong treatment for the majority of contemporary PDB patients who have disease of limited extent [
]. The less frequently encountered patients with early onset and more extensive disease are likely to require repeated treatment, but even in this group a zoledronate infusion should induce remission lasting several years.
In any individual, particularly if they have only one or two bones involved, the precise contribution of PDB to musculoskeletal symptoms can be difficult to ascertain. If there is a reasonable suspicion that PDB is causing pain, then most clinicians will recommend bisphosphonate treatment, taking the pragmatic view that if the pain resolves (and stays away) then Paget's disease was responsible. Overtreatment should be avoided and, if after treatment, pain persists or recurs early despite bone turnover being suppressed, there is nothing to be gained by administering more bisphosphonate.
8. What Sort of Disease is It?
While there have been many substantial advances in understanding the epidemiology, genetics and treatment, PDB it remains an enigmatic disease. It is a disorder which appears to be a unique to humans, and with few parallels to other human diseases. In some respects it behaves as a multifocal benign tumor. Although the clinical manifestations are osteoclast-led, anti-osteoclastic treatment with potent bisphosphonates do not permanently cure the disease, suggesting that a possibly non-osteoclastic ‘stem’ cell remains within bone; the transmission of PDB by bone grafting would support such a model. As PDB relapses after bisphosphonate treatment the rise in ALP, reflecting pagetic ‘mass’, follows the Gompertz curve, a common mathematical description of tumor growth [
]. But what is the nature of this ‘stem’ cell, and how might it recruit the osteoclasts that create the disease?
A perennial problem in explaining PDB is the apparently synchronous but random distribution of lesions within the skeleton, and the non-appearance of new lesions. This is a particular issue when an inherited genetic cause is invoked, as all cells within all bones will carry the germline mutation. One parallel here is metastatic disease in which somatic mutations are of the greatest significance. Somatic mutations accumulate as we age, so the increased prevalence of PDB with aging would fit in with this hypothesis. Apart from two studies that examined whether somatic mutations in SQSTM1 are present in pagetic bone (and reached opposing conclusions) [
], there has been little attention given to the possibility that somatic mutation might underlie PDB. The ‘metastatic’ model implies hematogenous spread, which would fit in with the random distribution of lesions. However, as no new lesions appear to develop, it suggests just a single episode of hematogenous spread – which would be most atypical of metastatic disease in general.
The characteristic osteoclastic inclusion bodies found in PDB were originally thought to represent viral particles. The Pittsburgh group have found evidence of measles virus nucleocapsid protein in pagetic osteoclasts and proposed a model whereby a virus might interact with genetic factors such as SQSTM1 mutations to produce the disease [
]. It has been proposed that the secular decline in PDB prevalence and severity might result from the introduction of measles immunization around 1971, but, as outlined above, the secular decline may well have begun before this time.
The changing epidemiology of PDB with later onset and less extensive disease, even in those with a genetic predisposition, argues that exposure to some predisposing environmental agent(s) has lessened over time. The lower disease burden in recently diagnosed patients also seems to imply a ‘dose effect’ that is difficult to explain, and the critical age for exposure is of course unknown. A number of epidemiological studies have linked PDB to rural life, exposure to various animals or to pollutants [
] (Table 2), but the nature of the environmental agent, and whether it is organic or inorganic remains unknown. Clearly there is still much to be learnt about, and from, Paget's disease.
Table 2Environmental exposures associated with Paget's disease of bone in epidemiological studies.
30182-8&id=gr1.jpg){kind=link}
30182-8&id=gr2.jpg){kind=link}
30182-8&id=gr3.jpg){kind=link}
30182-8&id=gr4.jpg){kind=link}