Fracture Risk in Monoclonal Gammopathy of Undetermined Significance


  • L Joseph Melton III,

    Corresponding author
    1. Division of Epidemiology, Department of Health Sciences Research, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
    2. Division of Endocrinology, Department of Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
    • Address reprint requests to: LJ Melton III, MD Division of Epidemiology Department of Health Sciences Research Mayo Clinic 200 First Street SW Rochester, MN 55905, USA
    Search for more papers by this author
  • S Vincent Rajkumar,

    1. Division of Hematology, Department of Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
    Search for more papers by this author
  • Sundeep Khosla,

    1. Division of Endocrinology, Department of Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
    Search for more papers by this author
  • Sara J Achenbach,

    1. Division of Biostatistics, Department of Health Sciences Research, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
    Search for more papers by this author
  • Ann L Oberg,

    1. Division of Biostatistics, Department of Health Sciences Research, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
    Search for more papers by this author
  • Robert A Kyle

    1. Division of Hematology, Department of Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, USA
    Search for more papers by this author

  • The authors have no conflict of interest.


To assess fractures in monoclonal gammopathy of undetermined significance (MGUS), the precursor of multiple myeloma, we followed 488 Olmsted County, MN, residents with MGUS in a retrospective cohort study. There was a 2.7-fold increase in the risk of axial fractures but no increase in limb fractures. The pathophysiologic basis for the increased axial fractures should be determined.

Introduction: Multiple myeloma is often preceded by monoclonal gammopathy of undetermined significance (MGUS). Fractures are common in myeloma as a result of lytic bone lesions, generalized bone loss, and elevated bone turnover from excessive cytokine production. Whether fractures are also increased in MGUS is unknown.

Materials and Methods: In a population-based retrospective cohort study, 488 Olmsted County, MN, residents with MGUS first diagnosed in 1960–1994 (52% men; mean age, 71.4 ± 12.8 years) were followed for 3901 person-years; follow-up was censored at progression to myeloma. The relative risk of fractures was assessed by standardized incidence ratios (SIRs), and risk factors were evaluated in proportional hazards models.

Results and Conclusions: Altogether, 200 patients experienced 385 fractures. Compared with expected rates in the community, statistically significant increases were seen for fractures at most axial sites, for example, vertebrae (SIR, 6.3; 95% CI, 5.2–7.5). There was a slight increase in hip (SIR, 1.6; 95% CI, 1.2–2.2) but not distal forearm fractures (SIR, 0.8; 95% CI, 0.4–1.5). The relative risk (SIR) of any axial fracture was 2.7 (95% CI, 2.3–3.1) compared with only 1.1 (95% CI, 0.9–1.4) for all limb fractures combined. In a multivariate analysis, the independent predictors of any subsequent fracture were age (hazard ratio [HR] per 10-year increase, 1.4; 95% CI, 1.2–1.6) and corticosteroid use (HR, 1.8; 95% CI, 1.2–2.6); greater weight at diagnosis (HR per 10 kg, 0.8; 95% CI, 0.8–0.9), and IgG monoclonal protein (HR, 0.7; 95% CI, 0.5–0.97) were protective. Baseline monoclonal protein level, a determinant of myeloma progression, did not predict fracture risk. Thus, the risk of axial, but not peripheral, fractures is increased among MGUS patients even before progression to myeloma. The pathophysiologic basis for this should be determined because elevated bone turnover, for example, might be treatable.


In 2003, nearly 15,000 Americans will develop multiple myeloma,(1) which has been recognized as an important cause of pathological fractures for over 100 years.(2) These fractures have typically been attributed to lytic lesions of bone, but patients can also have more generalized bone loss,(3–5) although little or no bone loss may be seen outside the lumbar spine in less severely affected patients.(6,7) It is unclear whether bone loss is also increased in patients with the premalignant precursor, monoclonal gammopathy of undetermined significance (MGUS). With progression from MGUS to myeloma, there is an increase in bone resorption(8,9) caused by the recruitment and activation of excessive numbers of osteoclasts by autocrine and paracrine cytokine production.(10,11) Some reports(9,12–14) but not others(8,15,16) indicate that elevated bone turnover may also be a feature of MGUS and may characterize the subset of MGUS cases that goes on to develop myeloma,(9) estimated at about 1% per year.(17) To test the hypothesis that the skeleton is unaffected in MGUS, we estimated the risk of fractures among Olmsted County, MN, residents with MGUS first recognized in 1960–1994. Because the MGUS population outnumbers myeloma by more than 20 to 1,(18) it is important to empirically quantify fracture risk in this group. If fracture risk were elevated because of excessive bone resorption, for example, this could likely be treated with one of the newer bisphosphonates.(19,20)


Olmsted County is well suited for disease association studies such as this because comprehensive medical records for the residents are available for review and are accessible through a centralized index to diagnoses made by essentially all medical care providers used by the local population.(21) After approval by Mayo's Institutional Review Board, we used this unique database (the Rochester Epidemiology Project) to identify the 1384 patients who resided in southeastern Minnesota when they were first diagnosed with MGUS in 1960–1994.(17) The present analysis was restricted to the subset of 514 patients who resided in Olmsted County, where fracture outcomes could be ascertained from review of comprehensive community medical records. On rechecking, however, 26 patients were not actually residents of Olmsted County at diagnosis. Thus, the analysis was based on 488 subjects with MGUS, all of whom had provided an authorization for review of their medical records for research in accordance with Minnesota privacy law.(22)

As described in detail elsewhere,(17) these patients had serum monoclonal protein concentrations of 3 g/dl or less. Monoclonal proteins were identified by cellulose acetate or agarose-gel electrophoresis; if there was an abnormal band or equivocal pattern, immunoelectrophoresis or immunofixation was performed. Bone marrow examination is not necessary in MGUS patients unless the monoclonal protein value is more than 2 g/dl or the patient has unexplained anemia, renal insufficiency, hypercalcemia, or bone pain; if bone marrow examination was performed, however, the plasma cell content had to be 10% or less. Thus, patients with smoldering multiple myeloma, characterized by a monoclonal protein value greater than 3 g/dl or more than 10% plasma cells in the bone marrow, were excluded. Patients were advised to undergo serum protein electrophoresis annually and were contacted if they did not.

After additional approval by Mayo's Institutional Review Board, these subjects were followed forward in time through their linked medical records in the community (retrospective cohort study) until death, development of myeloma or related disorder (IgM lymphoma, primary amyloidosis, macroglobulinemia, chronic lymphocytic leukemia, or plasmacytoma), or the most recent clinical contact. For each subject, all inpatient and outpatient records at any local provider of medical care were searched for the occurrence of any fracture. Mayo Clinic records, for example, contain the details of every inpatient hospitalization, every outpatient office or clinic visit, and all emergency room and nursing home care, as well as all laboratory results, all radiographic and pathology reports, including autopsies, and all correspondence with each patient.(21) The records contained the clinical history and the radiologist's report of each fracture, but the original radiographs were not available for review. Thus, the diagnosis of vertebral fracture was accepted based on a radiologist's report of compression or collapse of one or more thoracic or lumbar vertebrae. Ascertainment of clinically evident fractures is believed to be complete except for vertebral fractures, some of which are never diagnosed.(23)

The influence of MGUS on subsequent fracture risk was evaluated using three basic methods of analysis: in the primary analysis, we calculated standardized incidence ratios (SIRs), comparing the number of fractures that were observed at each skeletal site (based on the first fracture of a given type per person) to the number expected in this cohort during their follow-up in the community. Expected numbers of fractures were derived by applying age- and sex-specific incidence rates from the local population for these fractures(23–28) to the age- and sex-specific person-years of follow-up in the MGUS cohort. Calculations were made separately for the subset of moderate trauma fractures traditionally associated with osteoporosis (proximal femur [hip], thoracic or lumbar vertebrae [spine], and distal forearm); the later calculations used expected rates for moderate trauma fractures in the community.(28) 95% CI for the SIRs were calculated assuming that the expected rates are fixed and the observed fractures follow a Poisson distribution.(29)

In the second method of analysis, the cumulative incidence of new fractures (1 − probability of survival-free-of-fracture) was projected for up to 25 years after the initial diagnosis of MGUS using product-limit methods.(30) Death was treated as a competing event. Cumulative incidence curves were compared using the log-rank test statistic.(31) Kaplan-Meier methods were used to assess survival as well, and the observed and expected survival curves were also compared using the log-rank test statistic.

Finally, Cox proportional hazards models(32) were used to assess the impact of various covariates on subsequent fracture risk. Univariate relationships between the risk of specific fractures and each clinical characteristic under consideration were first assessed. Stepwise methods with forward selection and backward elimination were used to choose independent variables for the final models, and interactions among the significant main effect terms were assessed. The dependent variable was time until the first new fracture, and the independent variables were age, sex, and clinical characteristics. The various drug exposures were handled as time-dependent variables. For the final multiple models, as well as for the univariate models, the assumption of proportional hazards was examined and was not violated for the variables considered. All analyses were carried out in SAS (SAS Institute Inc., Cary, NC, USA).


Of 488 Olmsted County residents first diagnosed with MGUS during the 35-year period, 1960–1994, all but 2 were white, reflecting the racial composition of the community (98% white in 1980). Their mean ± SD age at diagnosis was 71.4 ± 12.8 years (median, 73.3 years; range, 24–96 years), and 255 (52%) were men compared with 233 women. These subjects were subsequently followed for a total of 3901 person-years (median, 7.2 years per subject). Survival was somewhat impaired in this cohort: at 25 years, 15% remained alive compared with an expected 17% (p < 0.001). Follow-up was complete to death in 67% of the subjects (median, 5.6 years of follow-up per subject), whereas the median duration of follow-up was 9.7 years among survivors.

During this period of observation, 200 subjects experienced 385 different fractures (Table 1). Altogether, 102 patients had a single fracture, whereas 49 had two fractures, 27 had three fractures, and 22 had four or more fractures. After 20 years of follow-up, an actuarially estimated 50% of the patients had experienced at least one new fracture compared with an expected 43% (p < 0.001) when death was considered a competing risk (Fig. 1). Fifty-nine of the fractures (15%) were caused by severe trauma (12 motor vehicle accidents, 23 falls from a height, and 24 miscellaneous causes, e.g., hand crushed in car door), but the majority of fractures (276, 72%) were attributed to minimal or moderate trauma (by convention, the equivalent of a fall from standing height or less; Table 1). Over two-thirds of the limb fractures (excluding hands and feet) were caused by a fall from standing height or less, whereas 83% of all vertebral fractures occurred “spontaneously” in the course of everyday activities. Eighteen fractures resulted from a specific pathological lesion (e.g., mostly metastatic lung or prostate cancer), and the etiology of the remaining 32 fractures was uncertain.

Table Table 1. Distribution of Fractures Among 488 Olmsted County, MN Residents After a Diagnosis of Monoclonal Gammopathy of Undetermined Significance (MGUS) in 1960–1994, by Fracture Site and Cause
original image
Figure FIG. 1..

Observed vs. expected cumulative incidence of fracture (p < 0.001) among 488 Olmsted County residents after a diagnosis of MGUS in 1960–1994. Note that death was considered a competing risk in this analysis.

Fracture data for specific skeletal sites are delineated in Table 2. Compared with expected rates among community residents generally, statistically significant increases were seen for most fractures of axial skeletal sites, particularly the vertebrae (SIR, 6.3; 95% CI, 5.2–7.5). The relative risk of any axial fracture was 2.7 (95% CI, 2.3–3.1) compared with only 1.1 (95% CI, 0.9–1.4) for all limb fractures combined. There was a slight increase in hip fracture risk (SIR, 1.6; 95% CI, 1.2–2.2) but no increase in distal forearm fractures (SIR, 0.8; 95% CI, 0.4–1.5), so the overall increase associated with an osteoporotic fracture (moderate trauma fracture of the hip, spine, or distal forearm ≥35 years of age: SIR, 2.5; 95% CI, 2.1–2.9) was accounted for by the excess vertebral fractures. The increase in overall fracture risk was statistically significant among men (SIR, 2.3; 95% CI, 1.8–2.8) and women (SIR, 1.6; 95% CI, 1.3–1.9) separately, and in both instances, was attributable to an increase in fractures of the axial skeleton.

Table Table 2. Fractures Observed (Obs)* Among 488 Olmsted County, MN Residents After a Diagnosis of Monoclonal Gammopathy of Undetermined Significance (MGUS) in 1960–1994 Compared With the Numbers Expected (Exp) and the Standardized Incidence Ratios (SIR), With 95% CI
original image

In a multivariate analysis, the only independent predictors of any subsequent fracture were age (hazard ratio [HR] per 10-year increase, 1.4; 95% CI, 1.2–1.6) and corticosteroid use (HR, 1.8; 95% CI, 1.2–2.6), whereas greater weight at diagnosis of MGUS was protective (HR per 10 kg, 0.8; 95% CI, 0.8–0.9) as was having an elevated monoclonal protein (M-protein) of the IgG type (HR, 0.7; 95% CI, 0.5–0.97). The total M-protein level at baseline, an important predictor of progression to myeloma, was not an independent predictor of fractures in the univariate analysis and was not associated with fracture risk after adjusting for age, weight, and corticosteroid use (HR, 1.3; 95% CI, 0.98–1.7). Likewise, the presence of underlying comorbid conditions linked with secondary osteoporosis (HR, 1.0; 95% CI, 0.8–1.4) or diseases associated with an increased risk of falling (HR, 1.2; 95% CI, 0.7–2.2) were not independent predictors of overall fracture risk in the adjusted analysis.

The reduced fracture risk associated with an IgG monoclonal protein was most pronounced for limb fractures (HR, 0.6; 95% CI, 0.4–0.96). Greater height was also protective (HR, 0.7; 95% CI, 0.6–0.9), whereas older age was linked to an increased risk of limb fractures generally (HR, 1.6; 95% CI, 1.3–2.0). Likewise, age was a risk factor for axial fractures (HR, 1.4; 95% CI, 1.2–1.6), along with corticosteroid use (HR, 1.8; 95% CI, 1.2–2.7) and personal history of a previous osteoporotic fracture (HR, 1.5; 95% CI, 1.02–2.1); greater body weight was again protective (HR, 0.8; 95% CI, 0.8–0.9). The M-protein level was not predictive of axial fractures, but patients whose serum light chain was of the lambda type were at lower risk (HR, 0.7; 95% CI, 0.5–0.96).


These data indicate that the risk of fractures of the axial skeleton, but not of the peripheral skeleton, is increased among MGUS patients even before the disease has progressed to myeloma. This is an important observation because bisphosphonate therapy for MGUS is not currently recommended in the absence of evidence of adverse skeletal effects in these patients.(33) An increase in fracture risk would not necessarily have been predicted from preliminary work showing that lumbar spine and total body bone density was not impaired but rather elevated in 18 patients with MGUS.(34) In another small series of 18 MGUS patients, however, bone density was substantially reduced at the hip and spine compared with young normal values.(14) In any event, bone density is not the sole determinant of fractures. Elevated bone turnover is also an independent predictor of fracture risk,(35) and a number of studies have demonstrated elevated bone resorption and/or reduced bone formation among patients with MGUS.(9,12,13,), 36

Although the mechanisms responsible for this effect are still obscure, we found a lower risk of fractures generally and limb fractures specifically among MGUS patients with an IgG monoclonal protein, which accounted for 70% of the MGUS patients.(17) In univariate analyses, both IgA (HR, 1.3; 95% CI, 0.9–1.9) in 13% of patients and IgM (HR, 1.5; 95% CI, 1.02–2.1) in 17% were associated with greater fracture risk compared with IgG (HR, 0.7; 95% C, 0.5–0.9). In myeloma, preliminary data suggest that patients with an IgG heavy chain have less bone loss than those with an IgA monoclonal protein, and it has been speculated that IgA-producing plasma cells generate more of the cytokines that adversely influence bone metabolism.(37) We also found a lower risk of axial fractures in the 36% of patients who had a lambda light chain. In the univariate analysis, a lambda light chain was associated with a lower risk of axial fractures (HR, 0.7; 95% CI, 0.5–1.02) than was the presence of a kappa light chain (HR, 1.3; 95% CI, 0.95–1.8) in most of the remaining patients. Although we have no explanation for this result either, lambda light chains have been associated with sclerotic rather than lytic bone lesions in myeloma.(38) Note, however, that these reductions were only relative. Compared with expected rates in the community, patients with an IgG heavy chain (SIR, 1.7; 95% CI, 1.4–2.0) or lambda light chain (SIR, 1.5; 95% CI, 1.2–2.0) were at increased risk of fractures generally.

Alternatively, the apparent increase in fracture risk might be accounted for by bias. Selection bias could play a role if a serious underlying disease, which itself was associated with fractures, led to the serum protein evaluation and a subsequent diagnosis of MGUS. However, diseases linked to secondary osteoporosis or to an increased risk of falls were not predictors of fracture risk in this analysis. Moreover, there was no increase in limb fractures, presumably because of the absence of hematopoietic marrow at peripheral sites in the skeleton.(5) On the other hand, an increased risk restricted to fractures of the spine and ribs might be explained by detection bias: to the extent that MGUS patients are more likely than community residents generally to have chest X-rays, there is the potential for increased detection of otherwise undiagnosed vertebral and rib fractures. This is a potential problem because the expected number of vertebral fractures in this analysis was based on the incidence of clinically recognized vertebral fractures, which represent only about one-half of the vertebral deformities that can be detected on X-rays of spine.(23) A more complete assessment of vertebral fracture occurrence can be obtained by using incidence rates that are derived from the prevalence of vertebral deformities.(39) These rates overestimate true vertebral fracture incidence, however, because some vertebral deformities do not represent actual clinical fractures. Nonetheless, if the latter rates are substituted in the analysis, overall vertebral fracture risk is still elevated (SIR, 1.8; 95% CI, 1.5–2.2).

It has long been recognized that age is a risk factor for fractures generally,(40) that greater weight or body mass index (BMI) is protective,(41) and that the occurrence of one osteoporotic fracture increases the likelihood of additional fractures.(42) Consequently, the present results are not surprising. Likewise, corticosteroid use has a strong relation with bone loss and fracture risk(43) and has previously been implicated in the bone loss seen in patients with multiple myeloma.(4) To our knowledge, however, no previous study has evaluated the effect of corticosteroid use on bone loss or fracture risk among patients with MGUS. Of note, corticosteroids were not used for MGUS but for other indications (e.g., temporal arteritis) in each instance. It is unlikely that subsequent treatment for corticosteroid-induced osteoporosis had much effect on these data because the study period mostly predated the availability of potent bisphosphonates. Only 16 patients had any exposure to a bone antiresorptive agent (bisphosphonate in 8, selective estrogen receptor modulator in 7, and calcitonin in 1), and just 8 started treatment after the initiation of corticosteroid therapy. Forty-three women had been on hormone therapy at one time or another, but fracture risk was not reduced among these women (HR, 1.2; 95% CI, 0.8–1.9). Similarly, there was no influence of chemotherapy or radiation treatment on these results because patients were censored in this analysis if they exhibited progression to myeloma or a related plasma cell disorder.

The present investigation has a number of strengths. These study subjects represented a large, population-based inception cohort that included both institutionalized and community-dwelling individuals registered at the time their MGUS was first recognized. Because of the unique records linkage system in Rochester, which provides access to the medical records of the entire community,(21) there should be nearly complete ascertainment of MGUS to the extent that the condition came to clinical attention.(17) Their clinical characteristics were recorded before any knowledge of the resulting fractures, which were documented in the detailed inpatient and outpatient medical records that spanned each subject's entire period of residency in the community. Because the vast majority of fractures come to medical attention,(28) ascertainment should be nearly complete with the possible exception of rib and vertebral fractures. Even so, the clinically detected vertebral fracture incidence rates in this community are much greater than those reported by others.(23) Also, there was considerable follow-up and a large number of fractures subsequent to the recognition of MGUS, which provided adequate statistical power.

There are also corresponding limitations of a study based on medical records. In particular, special measurements such as bone densitometry or biochemical markers of bone turnover were not routinely performed. A further limitation is the generalizability of these data from a small Midwestern community that is predominantly white and better educated than the white population of the United States as a whole.(21) Despite these limitations, this is the only population-based study that has estimated long-term fracture risk after an initial diagnosis of MGUS. Demonstration that the risk of axial fractures is substantially elevated in these patients indicates a need for further assessment of the pathophysiologic mechanisms that are responsible. Even before this issue is resolved, however, it may be appropriate to evaluate patients with MGUS for excessive bone loss or elevated bone turnover because both are clearly associated with increased fracture risk and both respond to antiresorptive therapy.(35)


The authors thank Leona Bellrichard, Joan LaPlante, Barbara Nolte, and Kristine Otto-Higgins for assistance with data collection and Mary Roberts for help in preparing the manuscript. This project was supported in part by grants AG-04875, CA-62242, and AR-30582 from the National Institutes of Health, U.S. Public Health Service.