This epidemiologic study of Paget's disease of bone used data from 788 cases and 387 spouse controls to investigate the following: (1) the extent to which this disorder aggregates in families; (2) the cumulative incidence of the disease in first-degree relatives of patients throughout life; and (3) the influence of age at diagnosis (<55 versus 55+ years) and presence of bone deformity in the case on risk of Paget's disease in relatives.
A positive family history in parents or siblings was reported by 12.3% of cases and 2.1% of controls. The rate of Paget's disease was approximately seven times as high in relatives of cases as in relatives of controls, and this increased rate did not differ according to gender of case or control or gender of relatives. Cumulative incidence of Paget's disease to age 90 was much higher in relatives of cases (8.9 ± 1.0% SEM) than in relatives of controls (1.8 ± 0.9% SEM). Among relatives of cases, cumulative risk was highest when the case had both early age at diagnosis and bone deformity (20.7 ± 3.6% SEM) compared with risk when the case had early age at diagnosis but not bone deformity (10.8 ± 3.2% SEM), bone deformity but not early age at diagnosis (5.8 ± 1.3% SEM), or neither bone deformity nor early age at diagnosis (3.6 ± 0.8% SEM). Risk in siblings of cases was higher when a parent was affected (22.1 ± 8.0% SEM) than when both parents were unaffected (6.7 ± 1.1% SEM).
These findings suggest that first-degree relatives of patients with Paget's disease have increased risk of developing the disorder, especially if the affected relatives have early age at diagnosis or deforming bone disease.
Paget's disease of bone is a common disorder of the skeleton believed to affect 1–3% of the population over age 40 and 8–10% of those over age 80 in the United States(1) Although the majority of patients are asymptomatic, a substantial number experience a variety of signs and symptoms, including bone pain, bone deformity, secondary arthritic complications, hearing loss, and syndromes of neurologic compression.
The pathophysiology of Paget's disease involves a localized disorder of bone remodeling, initiated by morphologically and functionally aberrant osteoclasts at affected skeletal sites. The etiology of the osteoclast lesion is unknown. A viral cause has been proposed based upon evidence of a paramyxoviral infection of these cells, but a specific virus has not yet been isolated.(2,3) Epidemiologic studies indicate that the condition is fairly common in some parts of the world (United Kingdom, United States, Canada, Australia, and New Zealand) and relatively rare in other areas (Scandinavia, Africa, and Asia).(1)
It has been known for many years that Paget's disease sometimes occurs in more than one member of a family. Pedigree studies led to the suggestion that in some unusual families susceptibility to the disorder may have an autosomal dominant mode of inheritance.(4) Very little is known, however, about the proportion of patients with a positive family history, or clinical differences between familial and nonfamilial Paget's disease.
This study used data collected from cases with Paget's disease and unaffected controls to address three questions about the role of familial aggregation in this disease. First, to what extent is the risk of Paget's disease increased in relatives of affected individuals? Second, in relatives of patients with Paget's disease, what is the cumulative risk of developing the disorder at specific ages throughout life? Third, to what extent is early age at diagnosis or disease severity in cases, as manifested by presence of bone deformity, related to the risk of Paget's disease in relatives?
MATERIALS AND METHODS
Subjects with Paget's disease (cases) were recruited through an announcement in the newsletter of the Paget's Disease Foundation in the United States. The announcement asked for volunteers to participate in a questionnaire study of the causes, natural history, clinical manifestations, and treatment of this disorder. No specific mention of family history was made in the announcement. Volunteers mailed back a short form from the newsletter requesting either a patient questionnaire only or both a patient and a control questionnaire for an unaffected spouse or, if there was no spouse, an unrelated friend of the same approximate age. Requests for questionnaires were received from 1056 patients, of whom 713 also requested control questionnaires. Cases and controls were mailed the same self-administered questionnaire, except that the version for controls omitted the questions that focused on pagetic signs, symptoms, and treatment. Completed questionnaires were returned by 864 (81.8%) cases and 503 (70.5%) controls.
Patients with Paget's disease were asked to report on age at diagnosis by a physician and presence and site(s) of bone deformity. Bone deformity was used as an indicator of disease severity in the cases.
For each biologic parent and full sibling, the questionnaire asked the relative's vital status, current age or age at death, and whether or not the case or control would be likely to know if the relative had a health problem. Relatives were excluded from the analysis if the subject was unable to report the relative's current age or age at death or reported that he or she would not be likely to know about the relative's health problems even if they were serious. For each relative the questionnaire also asked whether or not a doctor had diagnosed Paget's disease and at what age symptoms, if present, first appeared. If the question about Paget's disease in a relative was left blank, the relative was assumed to be unaffected; if the question was answered “don't know,” the relative was excluded from the analysis.
To minimize differences in awareness of Paget's disease in cases and controls, only spouse controls were included in the present study. A total of 387 cases had spouse controls, but the remaining 401 cases did not have spouse controls. Cases with and without controls were compared with respect to the proportion of their relatives who were reported to have Paget's disease. The proportion of affected relatives was similar in these two groups (cases with controls, 3.4%; cases without controls, 4.3%), suggesting that the two groups of cases were comparable with respect to the outcome of interest. Thus in the analysis cases with and without controls were combined and compared with the controls. The final study population comprised families of 788 cases and 387 controls, including 2930 relatives of cases (1299 parents and 1631 siblings) and 1360 relatives of controls (659 parents and 701 siblings).
Actuarial life table methods(5) were used to calculate cumulative incidence of Paget's disease in relatives of cases and controls, by age of relatives and age at diagnosis (< 55 versus ≥55 years) and bone deformity (present versus absent) in the cases. For this analysis relatives were assumed to be at risk of developing Paget's disease until their age at first symptom appearance (if affected) or current age or age at death (if unaffected). Since an age at first symptom appearance could not be obtained for relatives with asymptomatic Paget's disease, these relatives were assumed to have developed the disease at their current age or age at death. Equality of the cumulative incidence curves was tested using a generalized Savage statistic as proposed by Mantel.(6)
Cox proportional hazards regression models were used to calculate rate ratios for Paget's disease in relatives of cases versus relatives of controls. For this analysis maximum likelihood estimates of the regression coefficients and their standard errors were obtained by the method of Cox(7,8) using BMDP statistical software,(9) and the corresponding rate ratios and confidence intervals were obtained by exponentiation. These rate ratios are estimates of the ratio of age-specific incidence rates in relatives of cases versus relatives of controls, assuming that this ratio is constant with age. Thus they are adjusted for the years at risk of the relatives.
This study was approved by the Institutional Review Board of the Columbia-Presbyterian Medical Center.
A positive family history of Paget's disease in first-degree relatives (parents and full siblings) was reported by 97 (12.3%) of the 788 cases and 8 (2.1%) of the 387 controls. Among the cases with a positive family history 54 reported affected parents only (21 fathers and 33 mothers), 35 reported affected siblings only, and 8 reported both an affected parent and affected siblings (7 families in which the mother and one sibling were affected; 1 family in which the father and three siblings were affected). The majority of the families with only affected siblings contained only one affected sibling (29 families); in 5 additional families there were two affected, and in 1 family there were three affected siblings. None of the offspring was reported to be affected. All the 8 controls with a positive family history reported affected parents only (four fathers and four mothers); no affected siblings or offspring were reported.
Table 1 shows the crude percentage of affected relatives and rate ratio for Paget's disease in relatives of cases versus relatives of controls, by relationship (parent or sibling), gender of case or control, and gender of relative. Since none of the 701 control siblings was reported to be affected, the rate ratio for siblings could not be estimated. Overall the rate ratio was about 7, indicating that the rate of Paget's disease was about seven times as high in relatives of cases as in relatives of controls. There were no significant differences in the rate ratio according to either gender of case or control (6.8 versus 6.7, Table 1), or gender of the relatives (6.0 versus 7.5). Since the overall rate ratio was approximately 7, the lower rate ratio in parents (4.2, Table 1) suggests that the extent of increase in risk may be higher in siblings than in parents.
Figure 1 shows cumulative incidence by age for relatives of cases and relatives of controls, respectively. Cumulative incidence to age 90 was much higher in relatives of cases than in relatives of controls (8.9 ±1.0 versus 1.8 ± 0.9% SEM). Among relatives of cases there was little difference in risk between parents and siblings (cumulative incidence to age 90 in parents 8.4 ± 1.2% SEM, siblings 8.9 ± 1.7% SEM). There was also little difference in risk between male relatives (fathers and brothers) and female relatives (mothers and sisters) of cases (9.5 ± 1.7 versus 8.5 ± 1.3% SEM) or relatives of male and female cases (8.8 ± 1.5 versus 9.0 ± 1.5% SEM).
We also evaluated the relationship between risk of Paget's disease in relatives and age at diagnosis and presence of bone deformity in the cases. Cumulative incidence of Paget's disease to age 90 was higher in relatives of cases diagnosed before age 55 than in relatives of cases diagnosed at age 55 or older (7.0 ± 2.6 versus 4.6 ± 0.8% SEM; p < 0.0001). Risk to age 90 was also higher in relatives of cases with bone deformity than in relatives of cases without bone deformity (12.2 ± 1.7 versus 5.5 ± 1.1% SEM; p < 0.003). When these two variables were analyzed simultaneously (Fig. 2) cumulative incidence to age 90 was lowest (3.6 ± 0.9% SEM) in relatives of cases without bone deformity who had late age at diagnosis; this risk was still significantly higher (p < 0.002) than that found in relatives of the controls (1.8 ± 0.8% SEM). Risk was 5.6 ± 1.3% SEM in relatives of cases with bone deformity and late age at diagnosis and was 10.8 ± 3.2% SEM in relatives of cases without bone deformity who had early age at diagnosis. By far the highest risks were for relatives of cases with bone deformity and early age at diagnosis (20.7 ± 3.6% SEM). Bone deformity in the case was associated with increased risk among relatives of cases with both early and late age at diagnosis. Similarly, early age at diagnosis in the case was associated with increased risk among relatives of cases both with and without bone deformity.
One possible explanation for the association between age at diagnosis in the case and risk of Paget's disease in relatives is that awareness of Paget's disease in a family member causes individuals with symptoms possibly due to Paget's disease to seek medical attention at earlier ages. An affected parent is usually diagnosed before disease onset in the case, and thus presence of an affected parent may lead to diagnosis of the case at a younger age. On the other hand, diagnosis in a sibling is equally likely to precede or to follow that of the case. Presence of an affected sibling should therefore be less likely than presence of an affected parent to lead to younger age at diagnosis in the case.
To investigate the degree to which this potential detection bias might explain the association between age at diagnosis and risk in relatives, we examined this association in parents and siblings separately. The rate ratios in parents and siblings were nearly identical (2.49, 95% CI 1.49–4.17 versus 2.52, 95% CI 1.44–4.44); that is, in both parents and siblings the rate of Paget's disease was about 2.5 times as high in relatives of cases diagnosed before age 55 as in relatives of those diagnosed at age 55 or older. The similarity of this effect in parents and siblings suggests that it is not entirely accounted for by detection bias.
Table Table 1.. Percentage of Relatives with Paget's Disease and Rate Ratio for Paget's Disease Diagnosis in Relatives of Cases Versus Relatives of Controls, by Relationship, Gender of Case or Control, and Gender of Relative
Finally, risk in the siblings of cases was investigated in relation to the presence or absence of Paget's disease in a parent. Cumulative incidence to age 90 was higher among siblings of cases with an affected parent (N = 114, 22.1 ± 8.0% SEM) than among siblings of cases with unaffected parents (N = 1517, 6.7 ± 1.1% SEM). Among siblings of cases with an affected parent, risk was also higher when the mother was affected than when the father was affected (27.7 ± 12.4 versus 16.0 ± 10.4% SEM; p = 0.019).
This study investigated familial aggregation of Paget's disease of bone in a large population of patients who are members of a national voluntary organization. In this study population several important findings emerged. First, there was a substantial increase in risk of Paget's disease in first-degree relatives of cases compared with first-degree relatives of controls, and this increase in risk did not vary according to the gender of the case or the relative. Second, in this population the cumulative probability of a Paget's disease diagnosis by age 90 in a first-degree relative of an affected individual was approximately 9%, and this risk was the same regardless of the gender of the case, the gender of the relative, or the relationship (parent versus sibling) to the case. Third, among relatives of the cases certain characteristics of the cases were associated with substantial increases in familial risk: an age at diagnosis younger than 55 years, presence of bone deformity, and presence of an affected parent.
Generalizability of our findings may be limited by our use of a voluntary organization for case ascertainment. As we have previously noted,(10) individuals with severe or disabling disease may be more likely to join such organizations than less severely affected individuals. Further, since our findings indicate an association between family history and severity of Paget's disease, the risk estimates for relatives of cases in our study may be higher than for un-selected patients. This problem should not influence our internal comparison of cases with versus without bone deformity or early age at diagnosis, however.
Our estimates of cumulative incidence are Biased toward older ages, since many affected relatives were asymptomatic and we assumed they had disease onset at their current ages or ages at death. This would be a problem in any retrospective study of Paget's disease, since it is difficult to ascertain the time of onset of an asymptomatic disease. Further, even in symptomatic individuals the age at first symptom appearance is likely to be several years after disease onset.
Despite these limitations our findings aid in the assessment of risk in relatives of Paget's disease patients by providing estimates of cumulative incidence throughout life in these relatives (Figs. 1 and 2). Although the increase in risk to relatives of cases versus relatives of controls was estimated as sevenfold, the lifetime risk in relatives of cases was only 9%. This may be reassuring to relatives of patients. However, risks were substantially higher in relatives of cases diagnosed before age 55, especially if the case also had bone deformity.
We also found that risks in siblings of cases were higher if a parent was affected than if both parents were unaffected. This finding may be partly attributable to detection bias: siblings of affected individuals may be more likely to seek medical attention for mild or asymptomatic disease if a parent is also affected than if both parents are unaffected.
Our finding of higher risk in siblings when the mother was affected than when the father was affected is unexplained. This finding is not consistent with an X-linked genetic susceptibility, since there were no differences in risk between male and female relatives of cases. [Both X-linked dominant and X-linked recessive modes of inheritance predict large differences in risk to male versus female relatives.(11,12)] One interesting possibility is that there is a maternally transmitted influence on susceptibility, which could be mediated either by maternal environmental factors (such as a virus), cytoplasmic factors, or imprinting of a susceptibility gene.(11,13)
In our study population none of the cases had affected offspring, probably because the children of the cases were younger than the age at which the diagnosis is typically made. However, the risks in offspring are expected to be approximately equal to the risks in parents of cases (Fig. 1), assuming that there is no secular change in incidence.
Most patients with Paget's disease are asymptomatic, but a significant subset develop progressive and disabling complications.(14) Our finding of an association between severity of disease and familial risk (Fig. 2) suggests that patients with a positive family history may have higher susceptibility to symptomatic, deforming bone disease than patients with a negative family history.
Only one previous study has examined familial factors in Paget's disease in a large study population. In that investigation Sofaer ascertained patients from a voluntary organization in Scotland.(15) The proportion of cases with a positive family history was 13.8%, in agreement with our finding of 12.3%. Sofaer also reported a 10-fold increase in the prevalence of Paget's disease among parents and siblings of cases versus spousal controls, as well as earlier ages at first awareness of disease in cases with a positive family history. Sofaer suggested that risks were higher in female than in male relatives of cases, but this difference was not found in our study.
Our study contributes several methodologic advances over the previous work in this area. First, our analytic procedures included adjustment for the years at risk of the relatives. Second, we specifically asked cases and controls to assess their awareness of the health status of their relatives and included relatives only if the case or control indicated knowledge of their health histories. This procedure was included to improve the validity of the reported family history data. Finally, we provided estimates of cumulative risk of developing Paget's disease for relatives at specific ages throughout life.
We wish to emphasize that this study represents only a first step in understanding factors contributing to familial aggregation of Paget's disease. The observed familial aggregation could result either from genetic or nongenetic factors. The reported evidence supporting a viral etiology for this disease(2,3) suggests two possible mechanisms for the familial effects. First, family members sharing a common environment may be more likely than unrelated persons to have common exposure to an infectious agent or to transmit the agent to each other through personal contact. It has been shown that with this mechanism a pattern of maternal transmission, such as that we have observed, is expected.(16) Second, there may be an inherited susceptibility to infection by a relevant organism. Some preliminary studies have provided evidence for linkage of a Paget's disease susceptibility gene to HLA, possibly suggesting increased susceptibility of osteoclasts to viral infection.(17–19) In some families, genetic or familial environmental factors unrelated to viral infection may also be important. The contributions of these mechanisms to familial aggregation of Paget's disease should be explored in future studies.
This study was supported by a grant from the Paget's Disease Foundation and by NIH grants RR000645 and NS20656.
JBMR Anniversary Classic: Familial Aggregation of Paget's Disease of Bone
ES Siris, R Ottman, E Flaster, JL Kelsey
Originally published in Volume 6, Number 5 pp 495–500 (1991)
This survey of familial aggregation of Paget's disease stimulated the substantial advances in our understanding of genetic and environmental factors in Paget's disease that occurred during the next two decades.