The Spine Journal
-
(2015)
-
Clinical study
Radiologic and clinical characteristics of vertebral fractures in multiple myeloma Jacob A. Miller, BSa,b, Andrew Bowen, BSa,b, Megan V. Morisada, BSa,b, Konstantinos Margetis, MD, PhDa,c, Daniel Lubelski, BAa,b, Isador H. Lieberman, MD, MBAd, Edward C. Benzel, MDa,b,c, Thomas E. Mroz, MDa,b,* a
Cleveland Clinic Center for Spine Health, Cleveland Clinic, 9500 Euclid Ave., S-80, Cleveland, OH 44195, USA b Cleveland Clinic Lerner College of Medicine, 9500 Euclid Ave., NA-24, Cleveland, OH, 44195, USA c Department of Neurological Surgery, Cleveland Clinic, 9500 Euclid Ave., S-80, Cleveland, OH 44195, USA d Texas Back Institute, 6020 West Parker Rd. #200, Plano, TX, 75093, USA Received 10 January 2015; revised 27 March 2015; accepted 19 May 2015
Abstract
BACKGROUND CONTEXT: Nearly 80% of patients with newly diagnosed multiple myeloma (MM) have bony lesions on magnetic resonance imaging (MRI). These lesions may progress to debilitating vertebral fractures. No studies have quantitatively characterized these fractures or identified predictors of fracture burden and severity. PURPOSE: The purpose of this study was to characterize the clinical and radiologic features of these fractures and to identify independent predictors of fracture burden and severity. STUDY DESIGN/SETTING: : A consecutive retrospective chart review was conducted from January 2007 to December 2013 at a single tertiary-care institution. PATIENT SAMPLE: Patients with diagnoses of both MM and vertebral fracture were included in this study. Those with a history of non-MM vertebral fracture were excluded. OUTCOME MEASURES: The primary outcome measure was height loss of the fractured vertebral body, whereas secondary outcome measures included number of fractures and morphology. METHODS: Data were collected at fracture presentation. Radiologic data were obtained from T1-weighted MRI. Anterior, middle, and posterior vertebral body height losses were recorded, and a Genant grading was made. Multivariable Poisson and logistic regression were performed to identify predictors of fracture burden and severity. RESULTS: Among 50 patients presenting with vertebral fracture, 124 fractures were observed. The majority (76%) of these patients did not have a previous MM diagnosis. The most common presenting symptom was back pain (84%), followed by neurologic (54%) and constitutional (50%) symptoms. The mean anterior, middle, and posterior height losses of the fractured vertebral body were 30%, 37%, and 16%, respectively. Twenty percent of fractures were Genant Grade 1 (mild), whereas 32% and 48% were grades 2 (moderate) and 3 (severe). Fifty-five percent of fractures were biconcave, whereas 32% and 13% were wedge and crush fractures. Lower body mass index and albumin and increased myeloma protein, light chains, and creatinine predicted an increased number of fractures at presentation. Increased b2-microglobulin and creatinine predicted more severe vertebral fractures.
FDA device/drug status: Not applicable. Author disclosures: JAM: Nothing to disclose. AB: Nothing to disclose. MVM: Nothing to disclose. KM: Nothing to disclose. DL: Nothing to disclose. IHL: Stock Ownership: Merlot OrthopediX, Axiomed Spine Corp., Mazor Surgical, Crosstress Medical, Pearldiver, Inc.; Consulting: Merlot OrthopediX (B), Axiomed Spine Corp. (B), Mazor Surgical (B), Crosstress Medical (B), Pearldiver, Inc. (B). ECB: Royalties: Elsevier Pub (B), Thieme Pub (B); Stock Ownership: Axiomed, Depuy, Orthomems, Turning Point; Consulting: Axiomed; Speaking and/or Teaching Arrangements: Multiple (B); Trips/Travel: Multiple (B); Grants: OREF (F, Paid directly to institution), Rawlings (F, Paid directly to institution); http://dx.doi.org/10.1016/j.spinee.2015.05.026 1529-9430/Ó 2015 Elsevier Inc. All rights reserved.
TEM: Stock Ownership: PearlDiver, Inc. (B); Consulting: Globus Medical (B); Speaking and/or Teaching Arrangements: AO Spine (B). The disclosure key can be found on the Table of Contents and at www. TheSpineJournalOnline.com. Institutional review board (#13-1351) approval was obtained before initiation of the study. * Corresponding author. Department of Orthopedic Surgery, Neurological Institute, Cleveland Clinic Center for Spine Health, Cleveland Clinic, 9500 Euclid Ave., S-80, Cleveland, OH 44195, USA. Tel.: (216) 445-9232; fax: (216) 363-2040. E-mail address: [email protected] (T.E. Mroz)
2
J.A. Miller et al. / The Spine Journal
-
(2015)
-
CONCLUSIONS: In the present study, 124 fractures were observed among 50 patients. These fractures were generally severe, biconcave, and in the thoracic spine. Laboratory signs of advanced MM predict greater fracture burden and severity. In the future, monitoring of these predictors may raise suspicion for an MM-associated vertebral fracture. Ó 2015 Elsevier Inc. All rights reserved. Keywords:
Multiple myeloma; Vertebral fractures; Genant grading; Imaging; Height loss; Plasmacytoma
Introduction Multiple myeloma (MM) is a neoplasm of terminally differentiated plasma cells that occurs in 6 per 100,000 patients annually [1,2]. It is relatively uncommon, accounting for 1% of all malignancies and 13% of hematologic malignancies. As a systemic disease, patients with MM may present with signs and symptoms of renal failure, hypercalcemia, anemia, or infection. However, these patients most commonly present with bone pain in the back or ribs [1,2]. Eighty percent of patients with newly diagnosed MM have osteolytic bony lesions on imaging [1,2]. As a consequence of osteolysis, hypercalcemia and anemia may be observed with or without fracture in the affected bone. Bone loss in MM may not always be clinically observable, as approximately 50% of bone must be resorbed to result in radiologic changes [3]. As a result of bone loss, vertebral fractures are common (55%–70%) in MM and are generally classified as benign or malignant based on the presence of a focal mass on magnetic resonance imaging (MRI) [1,3,4]. These vertebral fractures may occur in patients with active treated and untreated diseases and in patients with biopsyconfirmed remission [1]. Patients with initial MRI evidence of diffuse marrow disease suffer a shorter fracture-free interval after MM diagnosis than patients with a solitary plasmacytoma [1,4]. Steroid use and hypercalcemia have been identified as risk factors for these fractures. Unfortunately, little data are available regarding specific radiologic characteristics of these fractures. Patients with MM-associated vertebral fractures suffer from poorer quality of life than patients without fracture [5–11]. These fractures may be accompanied by radicular or myelopathic symptoms, limiting mobility and everyday activities [12]. Numerous studies have investigated the beneficial role of kyphoplasty in these patients. Preoperatively, patients report poor Short-Form 36, Oswestry disability index, pain Visual Analog Scale, and Roland Disability Questionnaire scores, all of which improve with surgical intervention [5–11]. Despite the burden of these fractures on patient quality of life, radiologic characteristics have been poorly characterized in the literature. In addition, although steroid use and hypercalcemia have been identified as predictors of fracture risk in MM, no studies have confirmed whether serum levels of myeloma-specific laboratory markers (myeloma protein [M-protein] and light chains) are predictive of fracture burden or severity.
Investigating the frequency, location, characteristics, and predictors of these fractures may contribute to standardized screening, surveillance, and referral regimens for spine health in this population. In the present study, we sought to characterize these vertebral fractures radiologically and identify predictors of greater fracture burden and severity at presentation. Methods Patient selection A consecutive retrospective chart review of all patients with a diagnosis of MM and identified vertebral fracture between January 2007 and December 2013 at a single tertiarycare institution was conducted. Patients were excluded if they had a history of vertebral fracture or vertebral-deforming pathologies (including developmental deformities, Scheuermann disease, osteomalacia, and Paget disease). Data collection Demographic, medical history, surgical history, presenting symptoms, laboratory, and radiologic data were collected in a retrospective fashion from electronic medical records. For those patients with vertebral fracture as the presenting sign of myeloma, data were collected from the primary MM workup. For those patients with a previous diagnosis of MM, data were collected from outpatient visits closest to the time of the first vertebral fracture. C-reactive protein (CRP), erythrocyte sedimentation rate, b2-microglobulin (b2-MG), kappa/lambda light chains, M-protein, albumin, calcium, creatinine, and hemoglobin were recorded from blood samples at initial presentation with myeloma or fracture. Surgical pathology reports of bone marrow aspirate were used to assess plasma cell burden as a percentage of total cellularity. Patients were staged according to both the International Staging System (ISS) and the Durie-Salmon Staging System (DSS) [13,14]. The DSS was developed in the 1970s and uses hemoglobin, calcium, M-protein, and the presence of bony lesions on imaging to stage patients. This system has been used as an acceptable surrogate for plasma cell mass in the bone marrow. In contrast, the ISS, a newer system based on indicators of survival, uses b2-MG and albumin to stage patients, with median survival for stages 1, 2, and 3 of 62, 44, and 29 months, respectively [15].
J.A. Miller et al. / The Spine Journal
Radiologic measurements Radiologic measurements of midsagittal T1-weighted MR images were performed closest to primary diagnosis of vertebral fracture. The entire spine was examined radiologically for fractures. Patients unable to undergo MRI instead were subject to computed tomography. Anterior, middle, and posterior heights of vertebral bodies were measured in a quantitative fashion [12,16–19]. The fractured and immediately adjacent vertebrae were measured. Height loss was calculated with respect to the mean heights of the adjacent unfractured vertebrae. A semiquantitative Genant grade was made for each fracture (Grade 1, 20%–25% height loss; Grade 2, 26%–40% height loss; and Grade 3, greater than 40% height loss) [16]. Finally, fractures were classified as wedge, biconcave, or crush morphology [16–18]. Statistical analyses Data were analyzed using the JMP Pro 10 statistical software (SAS Institute, Inc., Cary, NC, USA, 2012) [20]. All observations were assumed to be independent. Poisson and logistic distributions best reflected the number of fractures at presentation and vertebral body height loss (bound from 0% to 100%), respectively. As such, simple and multivariable Poisson and logistic regression modeling were performed to identify predictors of the number of fractures at presentation and greatest degree of vertebral body height loss. Independent variables associated (p!.2) with dependent variables on single-variable modeling were included in the multivariable models, to avoid overfitting. All covariates were tested for collinearity (R2$0.8), of which no evidence was observed. All values of p less than .05 were considered statistically significant.
Results Fifty patients were included in the study, with an average age of 65 years and a mean body mass index (BMI) of 27 and 62% were female (Table 1). The majority (76%) of patients presented with vertebral fracture without a previous diagnosis of MM (Fig. 1). Of those with a previous diagnosis of MM, the median time to first fracture was 4 months. Back pain was the most common presenting symptom (84%), followed by neurologic (radiculopathy, 32% and weakness, 22%) and constitutional (fatigue/malaise, 32%; weight loss, 10%; and nausea/vomiting, 8%) symptoms. Among all patients, 12% had a previous nonvertebral pathologic fracture related to MM. Twelve percent of patients were taking bisphosphonates before vertebral fracture diagnosis. Laboratory and staging data are presented in Table 2. At the time of first vertebral fracture, mean CRP (1.22 mg/dL), erythrocyte sedimentation rate (66 mm/h), and b2-MG (4.57 mg/L) were elevated. Mean albumin (3.69 g/dL) was low-normal, and calcium (uncorrected and corrected
-
(2015)
-
3
Table 1 Patient characteristics at presentation for vertebral fracture Characteristics
Statistics
n F Age at Dx of MM Age at first vertebral fracture Vertebral fracture at time of MM Dx* Time to first fracture (mo)y BMI History smoking Aggregate pack-years History alcohol abuse Comorbidities Hypertension Diabetes CKD Stage $3 CHF CAD Osteopenia/osteoporosis Cancer other than MM Presenting symptoms Back pain Fatigue/malaise Radiculopathy UE/LE weakness Weight loss Nausea/vomiting Constipation Past radiation treatment for MM Past nonvertebral pathologic fracture Rib Humerus Scapula Pelvis
50 31 (62) 65612 65612 38 (76) 4 (1,12) 2766 16 (32) 6613 1 (2) 23 5 5 1 4 16 10
(46) (10) (10) (2) (8) (32) (2)
42 16 16 11 5 4 4 3 6 5 1 1 1
(84) (32) (32) (22) (10) (8) (8) (6) (12) (10) (2) (2) (2)
BMI, body mass index; CAD, coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; Dx, diagnosis; F, female; LE, lower extremity; MM, multiple myeloma; UE, upper extremity. Note: Values are presented as mean6standard deviation or n (%) unless otherwise specified. Subtotals may exceed 100% if patient had multiple comorbidities, presenting symptoms, and so forth. * Percent of patients with vertebral fracture as presenting sign of MM. y Median (interquartile range) number of months elapsed between diagnosis of MM and first vertebral fracture (excluding fractures as presenting sign of MM).
for albumin) was normal (9.71 mg/dL). Forty percent of patients were hypoalbumenic, 34% were hypercalcemic, and 70% were anemic at presentation. The mean plasma cell marrow burden was 45% of total cellularity. Via the ISS, 28% had Stage 1 disease, 28% had Stage 2 disease, and 44% had Stage 3 disease. Via the DSS, 64% had Stage 2 disease, and 36% had Stage 3 disease. The radiologic characteristics of these fractures are presented in Table 3. One hundred twenty-four vertebral fractures were observed among the 50 patients. Most patients (56%) presented with a single vertebral fracture, but multiple fractures were common (44%). The mean anterior, middle, and posterior vertebral body height losses were 30%, 37%, and 16%, respectively. Twenty percent of fractures were Genant Grade 1 (mild), whereas 32% were Grade 2 (moderate) and 48% were Grade 3 (severe).
4
J.A. Miller et al. / The Spine Journal
-
(2015)
-
Fig. 1. Patients free from fracture after diagnosis of multiple myeloma (MM). Proportion of patients free from fracture plotted against time (in months) after initial diagnosis and workup for MM.
The majority of fractures had a biconcave morphology (55%). Most of the fractures occurred in the thoracic spine, with the greatest number at T12; the distribution of fractures throughout the spine is presented in Fig. 2. The distributions of the number and severity of these fractures on a per-patient basis are presented in Fig. 3. By simple Poisson regression, seven independent variables were associated (p!.2) with an increased number of fractures: decreased BMI increased b2-MG, decreased
serum albumin, increased M-protein, increased light chains, increased creatinine, and Durie-Salmon Stage 3 (Table 4). After multivariable regression, decreased BMI (p!.01), decreased albumin (p5.02), increased M-protein (p5.01), increased light chains (p!.01), and increased creatinine (p!.01) remained significantly and independently predictive of an increased number of fractures at presentation. Furthermore, the magnitudes of effect of these variables were robust: per unit decrease in BMI, a patient could
Table 2 Laboratory data and MM staging
Table 3 Fracture characteristics
Characteristics Laboratory data CRP (mg/dL) ESR (mm/h) b2-MG (mg/L) Serum albumin (g/dL) Hypoalbumenic Serum M-protein (g/dL) Serum light chains (g/L) Serum calcium (mg/dL) Corrected serum calcium (mg/dL) Hypercalcemic Serum creatinine (mg/dL) Serum hemoglobin (g/dL) Anemic* Marrow plasma cell (%) ISS Stage 1 Stage 2 Stage 3 DSS Stage 1 Stage 2 Stage 3
Statistics 1.2262.14 66651 4.5763.06 3.6960.67 20 (40) 2.4561.86 0.88 9.6661.29 9.7161.85 17 (34) 1.0060.42 10.9762.47 35 (70) 45628 14 (28) 14 (28) 22 (44) 0 (0) 32 (64) 18 (36)
b2-MG, b2-microglobulin; CRP, C-reactive protein; DSS, DurieSalmon Staging System; ESR, erythrocyte sedimentation rate; ISS, International Staging System; M-protein, myeloma protein. Note: Values are presented as mean6standard deviation or n (%). * Gender adjusted.
Characteristics
Statistics
Number of fractures Number of patients with 1 fracture 2 fractures 3–5 fractures O5 fractures Percent vertebral body height loss* Anterior Middle Posterior Genant grade 1 2 3 Fracture morphology Wedge Biconcave Crush Location Cervical Thoracic Lumbar
124 28 9 9 4
(56) (18) (18) (8)
30621 37621 16618 25 (20) 39 (32) 60 (48) 40 (32) 68 (55) 16 (13) 1 (1) 83 (67) 40 (32)
Note: Values are presented as mean6standard deviation or n (%). * Height loss is relative to average of adjacent levels (eg, average of T4 and T6 heights for a T5 fracture). Anterior measurement is height of vertebral body at anterior edge on sagittal midline T1-weighted magnetic resonance imaging slice at midline. The middle and posterior measurements are in the middle and posterior aspects of the vertebral body, respectively.
J.A. Miller et al. / The Spine Journal
-
(2015)
-
5
Fig. 2. Vertebral fractures by location. Distribution of 124 observed fractures plotted with respect to vertebral bodies.
expect 0.1 additional fractures; per unit (grams per deciliter) decrease in albumin, a patient could expect 0.3 additional fractures; per unit (grams per deciliter) increase in M-protein, a patient could expect 0.1 additional fractures; per unit (grams per liter) increase in light chains, a patient could expect 0.2 additional fractures; and finally, per unit (milligrams per deciliter) increase in creatinine, a patient could expect nearly one additional fracture. Similarly, simple logistic regression was performed to identify predictors of vertebral body height loss. Four independent variables were associated (p!.2) with an increased degree of height loss: a history of smoking, increased CRP, increased b2-MG, and increased creatinine. After multivariable regression, increased b2-MG (p5.04) and increased creatinine (p5.03) remained significantly and independently predictive of greater vertebral body height loss. Per unit increase in b2-MG (milligrams per liter) and creatinine (milligrams per deciliter), a patient could expect 4% and 58% additional height loss of the fractured vertebra, respectively.
Discussion Fracture incidence, characteristics, and radiology Previous studies have characterized the incidence, associated symptoms, and marrow characteristics of vertebral fractures in MM. Melton et al. [21] conducted a population-based study of 165 patients with MM; they found that 134 (81%) of these patients experienced 225 fractures (118 vertebral). This study established a reliable incidence of fractures in an MM population. Also in the present study, we found that vertebral fractures occur at first presentation with MM in 76% of patients. Sparse radiologic data have been reported for MMassociated vertebral fractures. Lecouvet et al. [3] reported the appearance and distribution of vertebral fractures in 37 patients with newly diagnosed MM. Among 224 vertebral fractures, 149 appeared benign and 75 appeared malignant based on MRI findings at any time in the clinical course (initial presentation or follow-up). The majority of these fractures were in the thoracic spine, similar to what was observed in the
Fig. 3. Distribution of dependent variables. (Top) Distribution of the number of patients sustaining single or multiple vertebral fractures. (Bottom) Distribution of the number of patients sustaining various vertebral body height losses.
6
J.A. Miller et al. / The Spine Journal
Table 4 Multivariable regression analysis Multivariable regression No. of presenting fractures* Independent variable
Percent height lossy
b Coefficient p Value Odds ratio p Value
Age at fracture (y) Gender (F) History of smoking BMI L0.103 Hypertension CAD CHF Diabetes CKD Stage $3 Osteopenia/porosis Hx other cancer CRP (mg/dL) ESR (mm/hr) b2-MG (mg/L) 0.351 Serum albumin (g/dL) L0.307 Serum M-protein (g/dL) 0.132 Serum light chains (g/L) 0.173 Serum calcium (mg/dL) Serum creatinine (mg/dL) 0.925 Serum hemoglobin (g/dL) Marrow plasma cell (%) Intl. MM stage [3] DSS [3] 0.347
1.054
.50
!.01
-
(2015)
-
height loss (37%) occurred in the middle of the vertebrae. In the osteoporotic population, it appears that greater anterior height loss is more common, likely as a result of axial loading and kyphosis of the thoracic spine [18]. Although axial loading and spinal curvature are certainly factors in the development of myeloma-associated fractures, the greater prevalence of biconcave fractures may relate to local disease in the marrow promoting interluekin-6–mediated osteolysis and central collapse of the vertebrae [1]. Risk factors for fracture burden and severity
0.930
.20
.94 .02 .01 !.01
1.036
.04
!.01
1.581
.03
.18
BMI, body mass index; Bx, biopsy; CAD, coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; CRP, C-reactive protein; DSS, Durie-Salmon Staging System; ESR, erythrocyte sedimentation rate; F, female; Hx, history; Intl., international; b2-MG, b2microglobulin; MM, multiple myeloma; M-protein, myeloma protein. Note: Bold values are statistically significant (p!.05) * Dependent variable is number of vertebral fractures at presentation. y Dependent variable is greatest percent height loss of the fractured vertebra.
present study. However, this study included only patients with DSS Stage 3 disease and before 1997, before the widespread use of bisphosphonates and modern chemotherapy for the prevention of subsequent fractures [22,23]. In another study, Lecouvet et al. [4] reported the correlation between MRI bone marrow patterns with vertebral fracture risk. They found that four basic marrow patterns existed (A–D, from normal marrow to diffuse marrow involvement) and that patients with more benign-appearing marrow experienced longer fracture-free survival. The relative risks for subsequent fractures with patterns C/D with respect to A/B were between 6.2 and 11.0. More than 90% of patients with patterns C/D experienced fractures within 3 months of MM diagnosis, whereas patients with benign patterns (A/B) experienced fractures later, at a median of 12 months. In the present study, the vast majority (76%) of patients presented with fractures without a previous MM diagnosis. Of those with a previous MM diagnosis, the median time to fracture was 4 months (Fig. 1). In the present series, the majority (55%) of fractures were biconcave in nature, and as such, the greatest mean
In a retrospective population-based review of 145 patients diagnosed with MM from 1945 to 2001, Melton et al. [21] reported that the average MM patient was anemic (11.4 g/dL) and normocalcemic (9.4 g/dL), with slight renal insufficiency (creatinine, 1.2 mg/dL) and a 40% bone marrow plasma cell burden. These values are similar to those observed in the present study. Using these data, Melton et al. [21] performed a multivariable analysis and observed that serum calcium, corticosteroid use, and chemotherapy were risk factors for subsequent fractures of any location. However, no laboratory values other than serum calcium were significantly associated with fracture risk. Lecouvet et al. [4] performed a similar multivariable analysis attempting to identify predictors of fracture-free survival in 37 newly diagnosed DSS Stage 3 MM patients. It was observed that increased b2-MG and cumulative steroid intake predicted first and second fracture-free survival, respectively. The present study identified risk factors that expand on the reports of Melton et al. [21] and Lecouvet et al. [4]. Two outcome variables were used: the number of presenting vertebral fractures and greatest height loss of the fractured vertebral body (anterior, middle, or posterior). Because 76% of patients presented with fracture without a previous diagnosis of MM, the present study did not investigate risk factors for first fracture. Decreased BMI and albumin and increased M-protein, light chains, and creatinine were found to be statistically significant predictors of multiple fractures at presentation after adjusting for confounders using a Poisson model. The magnitude of effect associated with these variables was substantial: for example, a patient with a BMI of 23 may expect one additional fracture compared with a patient with a BMI of 33. The wasting of malignancy or the greater prevalence of osteoporosis in patients with low BMI may explain this finding. As such, BMI and low albumin seem to be crude measures of advanced disease, and, therefore, potentially greater number of fractures. In addition, increases in myeloma-specific laboratory values (M-protein and light chains) and renal insufficiency carried a substantial risk for multiple fractures at presentation. These results suggest that patients with laboratory signs of more advanced disease may necessitate earlier or more frequent imaging. The present study also investigated independent predictors of more severe fractures at presentation. Increased b2-MG and
J.A. Miller et al. / The Spine Journal
creatinine were statistically significant independent predictors after adjusting for confounders. This likewise suggests that laboratory markers of MM are acceptable surrogates for identifying severe fractures. The clinician/surgeon may suspect MM in patients with back pain, decreased BMI and albumin, and increased creatinine, thereby avoiding excess or unnecessary imaging and laboratory tests in all patients with MM. Given the results of the present study and data reported in the literature, it appears that diagnosing MM before first fracture is exceedingly challenging, as approximately 75% of patients present with fracture at diagnosis. Furthermore, myeloma is a heterogeneous disease with different rates of progression, and routine screening for MM in healthy individuals is not performed. As such, MM patients generally present after onset of constitutional symptoms or laboratory anomalies (anemia, renal insufficiency, and so forth), and a high clinical suspicion for the disease is required. Identifying MM earlier in the disease course may prevent these fractures, which require an unknown duration of osteolysis before occurring. Given the identified risk factors in the present study, it appears that a number of accessible and routine measures are independently associated with fracture number and severity, including BMI, serum albumin, and serum creatinine. Although the specificity of these three variables is quite low, it is possible to develop nomograms from such combined data to determine the pretest probability of a myeloma-associated fracture. With these proposed nomograms and a high suspicion for myeloma, primarycare providers may proceed to sensitive and specific assays, such as b2-MG, M-protein, or light chains, for early identification of patients with myeloma. Alternatively, simple radiographs of the spine may be performed for fracture screening in patients with back pain demonstrating low BMI, low albumin, and elevated creatinine. Although initial vertebral fractures may not be preventable in all MM patients, earlier referral to spine clinics may lessen the morbidity associated with these fractures. Concomitantly, earlier use of bisphosphonates or RANK ligand inhibitors may serve to prevent or delay the progression of these fractures. Moreover, with available tools such as the Spinal Instability Neoplastic Score, clinicians are identifying patients at greatest risk for impending vertebral fracture [24]. In the myeloma population, a combined approach using dual-energy X-ray absorptiometry scores, Spinal Instability Neoplastic Score, and patient BMI and laboratory data is certainly feasible and practical, as these data are routinely collected. With this knowledge, prophylactic stabilization with established techniques such as kyphoplasty may alleviate bony morbidity before clinical deterioration. Limitations The present study had several limitations. Because of the retrospective nature of the investigation, data reported by clinicians were not standardized and subject to bias. Furthermore, this study was not a cohort study, and, therefore, did not
-
(2015)
-
7
follow a population of MM patients until first vertebral fracture. However, our data suggest that patients with MM are rarely diagnosed before development of a first vertebral fracture. It is also possible that there were patients who did not receive regular bone surveys; therefore, fractures were not identified in others who died before fracture detection. As such, a subset of undiagnosed fractures may not be characterized in this study. Last, although many variables were controlled for in the multivariable logistic regression, there may be unmeasured variables that confound the results. However, this study included variables related to demographics, comorbidities, laboratory values, and staging in an effort to identify any contributing confounding variables. The independent predictors identified in this study were not only statistically significant but biologically and clinically relevant in their relation to outcome variables.
Conclusions Vertebral fractures in MM are significantly detrimental to patient quality of life in a disease with a 45% five-year survival. Despite the burden of these fractures, they have been poorly characterized radiologically, and few independent risk factors for fracture burden have been established. In the present study, vertebral fractures associated with MM were generally severe, biconcave in morphology, and in the thoracic spine. Furthermore, in almost all patients, vertebral fracture was identified at the time of initial MM diagnosis. Given this, it is imperative to identify fractures early in their natural history such that medical and surgical interventions may be conducted. To aid in this objective, a number of independent risk factors for fracture burden and severity at presentation were identified, including lower BMI and albumin and higher b2-MG, M-protein, light chains, and creatinine. Based on these findings, it may be feasible to use nomograms of BMI, albumin, and creatinine to determine which patients in the primary-care setting are at greatest risk for MM. Subsequent laboratory testing with specific assays such as b2-MG may then diagnose myeloma and associated fractures earlier in the disease course. Future studies observing the progression of these fractures over clinical follow-up are warranted to characterize their rate of change and to establish evidence-based guidelines for spine surveillance in the myeloma population. References [1] Angtuaco EJC, Fassas ABT, Walker R, Sethi R, Barlogie B. Multiple myeloma: clinical review and diagnostic imaging. Radiology 2004;231:11–23. [2] Palumbo A, Anderson K. Multiple myeloma. N Engl J Med 2011;364:1046–60. [3] Lecouvet FE, Vande Berg BC, Maldague BE, Michaux L, Laterre E, Michaux JL, et al. Vertebral compression fractures in multiple myeloma. Part I. Distribution and appearance at MR imaging. Radiology 1997;204:195–9.
8
J.A. Miller et al. / The Spine Journal
[4] Lecouvet FE, Malghem J, Michaux L, Michaux JL, Lehmann F, Maldague BE, et al. Vertebral compression fractures in multiple myeloma. Part II. Assessment of fracture risk with MR imaging of spinal bone marrow. Radiology 1997;204:201–5. [5] Erdem E, Akdol S, Amole A, Fryar K, Eberle RW. Radiofrequencytargeted vertebral augmentation for the treatment of vertebral compression fractures as a result of multiple myeloma. Spine 2013;38:1275–81. [6] Dudeney S, Lieberman IH, Reinhardt M-K, Hussein M. Kyphoplasty in the treatment of osteolytic vertebral compression fractures as a result of multiple myeloma. J Clin Oncol 2002;20:2382–7. [7] Mendoza TR, Koyyalagunta D, Burton AW, Thomas SK, Phan MH, Giralt SA, et al. Changes in pain and other symptoms in patients with painful multiple myeloma-related vertebral fracture treated with kyphoplasty or vertebroplasty. J Pain 2012;13:564–70. [8] Julka A, Tolhurst SR, Srinivasan RC, Graziano GP. Functional outcomes and height restoration for patients with multiple myeloma-related osteolytic vertebral compression fractures treated with kyphoplasty. J. Spinal Disord Tech 2014;27:342–6. [9] Pflugmacher R, Kandziora F, Schroeder RJ, Melcher I, Haas NP, Klostermann CK. Percutaneous balloon kyphoplasty in the treatment of pathological vertebral body fracture and deformity in multiple myeloma: a one-year follow-up. Acta Radiol 2006;47:369–76. [10] McDonald RJ, Trout AT, Gray LA, Dispenzieri A, Thielen KR, Kallmes DF. Vertebroplasty in multiple myeloma: outcomes in a large patient series. AJNR Am J Neuroradiol 2008;29:642–8. [11] Berenson J, Pflugmacher R, Jarzem P, Zonder J, Sjchechtman K, Tillman JB, et al. Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body compression fractures in patients with cancer: a multicentre, randomised controlled trial. Lancet Oncol 2011;12:225–35. [12] Griffith JF, Adams JE, Genant HK. Diagnosis and classification of vertebral fracture. In: Rosen CJ, ed. Primer on the metabolic bone diseases and disorders of mineral metabolism. John Wiley & Sons, Inc., 2013: 317–35. Available at: http://onlinelibrary.wiley.com/doi/10.1002/ 9781118453926.ch37/summary. Accessed September 18, 2014. [13] Fechtner K, Hillengass J, Delorme S, Heiss C, Neben K, Goldschmidt H, et al. Staging monoclonal plasma cell disease:
[14]
[15] [16]
[17]
[18]
[19]
[20] [21]
[22] [23]
[24]
-
(2015)
-
comparison of the Durie-Salmon and the Durie-Salmon PLUS staging systems. Radiology 2010;257:195–204. Hari PN, Zhang M-J, Roy V, Perez WS, Bashey A, To LB, et al. Is the International Staging System superior to the Durie-Salmon staging system? A comparison in multiple myeloma patients undergoing autologous transplant. Leukemia 2009;23:1528–34. Anderson KC. Multiple myeloma: a clinical overview. Oncology (Williston Park) 2011;25(Suppl 2):3–9. Adams JE, Lenchik L, Roux C, Genant HK. Vertebral fracture initiative: radiologic assessment of vertebral fracture. Available at: http:// www.iofbonehealth.org/sites/default/files/PDFs/Vertebral%20Fractur e%20Initiative/IOF_VFI8Part_II-Manuscript.pdf. Accessed September 18, 2014. Grados F, Fechtenbaum J, Flipon E, Kolta S, Roux C, Fardellone P. Radiographic methods for evaluating osteoporotic vertebral fractures. Joint Bone Spine 2009;76:241–7. Jarraya M, Hayashi D, Griffith JF, Guermazi A, Genant HK. Identification of vertebral fractures. In: Guglielmi G, ed. Osteoporosis and bone densitometry measurements. Medical radiology. Berlin, Germany: Springer, 2013:41–55. Available at: http://link.springer.com/ chapter/10.1007/174_2012_701. Accessed September 18, 2014. Vaccaro AR, Oner C, Kepler CK, Dvorak M, Schnake K, Bellabarba C, et al. AOSpine thoracolumbar spine injury classification system: fracture description, neurological status, and key modifiers. Spine 2013;38:2028–37. JMP statistical software. Cary, NC: SAS Institute, Inc., 1989. Melton LJ III, Kyle RA, Achenbach SJ, Oberg AL, Rajkumar SV. Fracture risk with multiple myeloma: a population-based study. J Bone Miner Res 2005;20:487–93. Berenson JR, Lipton A. Bisphosphonates in the treatment of malignant bone disease. Annu Rev Med 1999;50:237–48. Berenson JR, Lipton A. Use of bisphosphonates in patients with metastatic bone disease. Oncology (Williston Park) 1998;12: 1573–9; discussion 1579–1581. Fourney DR, Frangou EM, Rkyen TC, Dipaola CP, Shaffrey CI, Berven SH, et al. Spinal instability neoplastic score: an analysis of reliability and validity from the spine oncology study group. J Clin Oncol 2011;29:3072–7.
-
(2015)
-
Clinical study
Radiologic and clinical characteristics of vertebral fractures in multiple myeloma Jacob A. Miller, BSa,b, Andrew Bowen, BSa,b, Megan V. Morisada, BSa,b, Konstantinos Margetis, MD, PhDa,c, Daniel Lubelski, BAa,b, Isador H. Lieberman, MD, MBAd, Edward C. Benzel, MDa,b,c, Thomas E. Mroz, MDa,b,* a
Cleveland Clinic Center for Spine Health, Cleveland Clinic, 9500 Euclid Ave., S-80, Cleveland, OH 44195, USA b Cleveland Clinic Lerner College of Medicine, 9500 Euclid Ave., NA-24, Cleveland, OH, 44195, USA c Department of Neurological Surgery, Cleveland Clinic, 9500 Euclid Ave., S-80, Cleveland, OH 44195, USA d Texas Back Institute, 6020 West Parker Rd. #200, Plano, TX, 75093, USA Received 10 January 2015; revised 27 March 2015; accepted 19 May 2015
Abstract
BACKGROUND CONTEXT: Nearly 80% of patients with newly diagnosed multiple myeloma (MM) have bony lesions on magnetic resonance imaging (MRI). These lesions may progress to debilitating vertebral fractures. No studies have quantitatively characterized these fractures or identified predictors of fracture burden and severity. PURPOSE: The purpose of this study was to characterize the clinical and radiologic features of these fractures and to identify independent predictors of fracture burden and severity. STUDY DESIGN/SETTING: : A consecutive retrospective chart review was conducted from January 2007 to December 2013 at a single tertiary-care institution. PATIENT SAMPLE: Patients with diagnoses of both MM and vertebral fracture were included in this study. Those with a history of non-MM vertebral fracture were excluded. OUTCOME MEASURES: The primary outcome measure was height loss of the fractured vertebral body, whereas secondary outcome measures included number of fractures and morphology. METHODS: Data were collected at fracture presentation. Radiologic data were obtained from T1-weighted MRI. Anterior, middle, and posterior vertebral body height losses were recorded, and a Genant grading was made. Multivariable Poisson and logistic regression were performed to identify predictors of fracture burden and severity. RESULTS: Among 50 patients presenting with vertebral fracture, 124 fractures were observed. The majority (76%) of these patients did not have a previous MM diagnosis. The most common presenting symptom was back pain (84%), followed by neurologic (54%) and constitutional (50%) symptoms. The mean anterior, middle, and posterior height losses of the fractured vertebral body were 30%, 37%, and 16%, respectively. Twenty percent of fractures were Genant Grade 1 (mild), whereas 32% and 48% were grades 2 (moderate) and 3 (severe). Fifty-five percent of fractures were biconcave, whereas 32% and 13% were wedge and crush fractures. Lower body mass index and albumin and increased myeloma protein, light chains, and creatinine predicted an increased number of fractures at presentation. Increased b2-microglobulin and creatinine predicted more severe vertebral fractures.
FDA device/drug status: Not applicable. Author disclosures: JAM: Nothing to disclose. AB: Nothing to disclose. MVM: Nothing to disclose. KM: Nothing to disclose. DL: Nothing to disclose. IHL: Stock Ownership: Merlot OrthopediX, Axiomed Spine Corp., Mazor Surgical, Crosstress Medical, Pearldiver, Inc.; Consulting: Merlot OrthopediX (B), Axiomed Spine Corp. (B), Mazor Surgical (B), Crosstress Medical (B), Pearldiver, Inc. (B). ECB: Royalties: Elsevier Pub (B), Thieme Pub (B); Stock Ownership: Axiomed, Depuy, Orthomems, Turning Point; Consulting: Axiomed; Speaking and/or Teaching Arrangements: Multiple (B); Trips/Travel: Multiple (B); Grants: OREF (F, Paid directly to institution), Rawlings (F, Paid directly to institution); http://dx.doi.org/10.1016/j.spinee.2015.05.026 1529-9430/Ó 2015 Elsevier Inc. All rights reserved.
TEM: Stock Ownership: PearlDiver, Inc. (B); Consulting: Globus Medical (B); Speaking and/or Teaching Arrangements: AO Spine (B). The disclosure key can be found on the Table of Contents and at www. TheSpineJournalOnline.com. Institutional review board (#13-1351) approval was obtained before initiation of the study. * Corresponding author. Department of Orthopedic Surgery, Neurological Institute, Cleveland Clinic Center for Spine Health, Cleveland Clinic, 9500 Euclid Ave., S-80, Cleveland, OH 44195, USA. Tel.: (216) 445-9232; fax: (216) 363-2040. E-mail address: [email protected] (T.E. Mroz)
2
J.A. Miller et al. / The Spine Journal
-
(2015)
-
CONCLUSIONS: In the present study, 124 fractures were observed among 50 patients. These fractures were generally severe, biconcave, and in the thoracic spine. Laboratory signs of advanced MM predict greater fracture burden and severity. In the future, monitoring of these predictors may raise suspicion for an MM-associated vertebral fracture. Ó 2015 Elsevier Inc. All rights reserved. Keywords:
Multiple myeloma; Vertebral fractures; Genant grading; Imaging; Height loss; Plasmacytoma
Introduction Multiple myeloma (MM) is a neoplasm of terminally differentiated plasma cells that occurs in 6 per 100,000 patients annually [1,2]. It is relatively uncommon, accounting for 1% of all malignancies and 13% of hematologic malignancies. As a systemic disease, patients with MM may present with signs and symptoms of renal failure, hypercalcemia, anemia, or infection. However, these patients most commonly present with bone pain in the back or ribs [1,2]. Eighty percent of patients with newly diagnosed MM have osteolytic bony lesions on imaging [1,2]. As a consequence of osteolysis, hypercalcemia and anemia may be observed with or without fracture in the affected bone. Bone loss in MM may not always be clinically observable, as approximately 50% of bone must be resorbed to result in radiologic changes [3]. As a result of bone loss, vertebral fractures are common (55%–70%) in MM and are generally classified as benign or malignant based on the presence of a focal mass on magnetic resonance imaging (MRI) [1,3,4]. These vertebral fractures may occur in patients with active treated and untreated diseases and in patients with biopsyconfirmed remission [1]. Patients with initial MRI evidence of diffuse marrow disease suffer a shorter fracture-free interval after MM diagnosis than patients with a solitary plasmacytoma [1,4]. Steroid use and hypercalcemia have been identified as risk factors for these fractures. Unfortunately, little data are available regarding specific radiologic characteristics of these fractures. Patients with MM-associated vertebral fractures suffer from poorer quality of life than patients without fracture [5–11]. These fractures may be accompanied by radicular or myelopathic symptoms, limiting mobility and everyday activities [12]. Numerous studies have investigated the beneficial role of kyphoplasty in these patients. Preoperatively, patients report poor Short-Form 36, Oswestry disability index, pain Visual Analog Scale, and Roland Disability Questionnaire scores, all of which improve with surgical intervention [5–11]. Despite the burden of these fractures on patient quality of life, radiologic characteristics have been poorly characterized in the literature. In addition, although steroid use and hypercalcemia have been identified as predictors of fracture risk in MM, no studies have confirmed whether serum levels of myeloma-specific laboratory markers (myeloma protein [M-protein] and light chains) are predictive of fracture burden or severity.
Investigating the frequency, location, characteristics, and predictors of these fractures may contribute to standardized screening, surveillance, and referral regimens for spine health in this population. In the present study, we sought to characterize these vertebral fractures radiologically and identify predictors of greater fracture burden and severity at presentation. Methods Patient selection A consecutive retrospective chart review of all patients with a diagnosis of MM and identified vertebral fracture between January 2007 and December 2013 at a single tertiarycare institution was conducted. Patients were excluded if they had a history of vertebral fracture or vertebral-deforming pathologies (including developmental deformities, Scheuermann disease, osteomalacia, and Paget disease). Data collection Demographic, medical history, surgical history, presenting symptoms, laboratory, and radiologic data were collected in a retrospective fashion from electronic medical records. For those patients with vertebral fracture as the presenting sign of myeloma, data were collected from the primary MM workup. For those patients with a previous diagnosis of MM, data were collected from outpatient visits closest to the time of the first vertebral fracture. C-reactive protein (CRP), erythrocyte sedimentation rate, b2-microglobulin (b2-MG), kappa/lambda light chains, M-protein, albumin, calcium, creatinine, and hemoglobin were recorded from blood samples at initial presentation with myeloma or fracture. Surgical pathology reports of bone marrow aspirate were used to assess plasma cell burden as a percentage of total cellularity. Patients were staged according to both the International Staging System (ISS) and the Durie-Salmon Staging System (DSS) [13,14]. The DSS was developed in the 1970s and uses hemoglobin, calcium, M-protein, and the presence of bony lesions on imaging to stage patients. This system has been used as an acceptable surrogate for plasma cell mass in the bone marrow. In contrast, the ISS, a newer system based on indicators of survival, uses b2-MG and albumin to stage patients, with median survival for stages 1, 2, and 3 of 62, 44, and 29 months, respectively [15].
J.A. Miller et al. / The Spine Journal
Radiologic measurements Radiologic measurements of midsagittal T1-weighted MR images were performed closest to primary diagnosis of vertebral fracture. The entire spine was examined radiologically for fractures. Patients unable to undergo MRI instead were subject to computed tomography. Anterior, middle, and posterior heights of vertebral bodies were measured in a quantitative fashion [12,16–19]. The fractured and immediately adjacent vertebrae were measured. Height loss was calculated with respect to the mean heights of the adjacent unfractured vertebrae. A semiquantitative Genant grade was made for each fracture (Grade 1, 20%–25% height loss; Grade 2, 26%–40% height loss; and Grade 3, greater than 40% height loss) [16]. Finally, fractures were classified as wedge, biconcave, or crush morphology [16–18]. Statistical analyses Data were analyzed using the JMP Pro 10 statistical software (SAS Institute, Inc., Cary, NC, USA, 2012) [20]. All observations were assumed to be independent. Poisson and logistic distributions best reflected the number of fractures at presentation and vertebral body height loss (bound from 0% to 100%), respectively. As such, simple and multivariable Poisson and logistic regression modeling were performed to identify predictors of the number of fractures at presentation and greatest degree of vertebral body height loss. Independent variables associated (p!.2) with dependent variables on single-variable modeling were included in the multivariable models, to avoid overfitting. All covariates were tested for collinearity (R2$0.8), of which no evidence was observed. All values of p less than .05 were considered statistically significant.
Results Fifty patients were included in the study, with an average age of 65 years and a mean body mass index (BMI) of 27 and 62% were female (Table 1). The majority (76%) of patients presented with vertebral fracture without a previous diagnosis of MM (Fig. 1). Of those with a previous diagnosis of MM, the median time to first fracture was 4 months. Back pain was the most common presenting symptom (84%), followed by neurologic (radiculopathy, 32% and weakness, 22%) and constitutional (fatigue/malaise, 32%; weight loss, 10%; and nausea/vomiting, 8%) symptoms. Among all patients, 12% had a previous nonvertebral pathologic fracture related to MM. Twelve percent of patients were taking bisphosphonates before vertebral fracture diagnosis. Laboratory and staging data are presented in Table 2. At the time of first vertebral fracture, mean CRP (1.22 mg/dL), erythrocyte sedimentation rate (66 mm/h), and b2-MG (4.57 mg/L) were elevated. Mean albumin (3.69 g/dL) was low-normal, and calcium (uncorrected and corrected
-
(2015)
-
3
Table 1 Patient characteristics at presentation for vertebral fracture Characteristics
Statistics
n F Age at Dx of MM Age at first vertebral fracture Vertebral fracture at time of MM Dx* Time to first fracture (mo)y BMI History smoking Aggregate pack-years History alcohol abuse Comorbidities Hypertension Diabetes CKD Stage $3 CHF CAD Osteopenia/osteoporosis Cancer other than MM Presenting symptoms Back pain Fatigue/malaise Radiculopathy UE/LE weakness Weight loss Nausea/vomiting Constipation Past radiation treatment for MM Past nonvertebral pathologic fracture Rib Humerus Scapula Pelvis
50 31 (62) 65612 65612 38 (76) 4 (1,12) 2766 16 (32) 6613 1 (2) 23 5 5 1 4 16 10
(46) (10) (10) (2) (8) (32) (2)
42 16 16 11 5 4 4 3 6 5 1 1 1
(84) (32) (32) (22) (10) (8) (8) (6) (12) (10) (2) (2) (2)
BMI, body mass index; CAD, coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; Dx, diagnosis; F, female; LE, lower extremity; MM, multiple myeloma; UE, upper extremity. Note: Values are presented as mean6standard deviation or n (%) unless otherwise specified. Subtotals may exceed 100% if patient had multiple comorbidities, presenting symptoms, and so forth. * Percent of patients with vertebral fracture as presenting sign of MM. y Median (interquartile range) number of months elapsed between diagnosis of MM and first vertebral fracture (excluding fractures as presenting sign of MM).
for albumin) was normal (9.71 mg/dL). Forty percent of patients were hypoalbumenic, 34% were hypercalcemic, and 70% were anemic at presentation. The mean plasma cell marrow burden was 45% of total cellularity. Via the ISS, 28% had Stage 1 disease, 28% had Stage 2 disease, and 44% had Stage 3 disease. Via the DSS, 64% had Stage 2 disease, and 36% had Stage 3 disease. The radiologic characteristics of these fractures are presented in Table 3. One hundred twenty-four vertebral fractures were observed among the 50 patients. Most patients (56%) presented with a single vertebral fracture, but multiple fractures were common (44%). The mean anterior, middle, and posterior vertebral body height losses were 30%, 37%, and 16%, respectively. Twenty percent of fractures were Genant Grade 1 (mild), whereas 32% were Grade 2 (moderate) and 48% were Grade 3 (severe).
4
J.A. Miller et al. / The Spine Journal
-
(2015)
-
Fig. 1. Patients free from fracture after diagnosis of multiple myeloma (MM). Proportion of patients free from fracture plotted against time (in months) after initial diagnosis and workup for MM.
The majority of fractures had a biconcave morphology (55%). Most of the fractures occurred in the thoracic spine, with the greatest number at T12; the distribution of fractures throughout the spine is presented in Fig. 2. The distributions of the number and severity of these fractures on a per-patient basis are presented in Fig. 3. By simple Poisson regression, seven independent variables were associated (p!.2) with an increased number of fractures: decreased BMI increased b2-MG, decreased
serum albumin, increased M-protein, increased light chains, increased creatinine, and Durie-Salmon Stage 3 (Table 4). After multivariable regression, decreased BMI (p!.01), decreased albumin (p5.02), increased M-protein (p5.01), increased light chains (p!.01), and increased creatinine (p!.01) remained significantly and independently predictive of an increased number of fractures at presentation. Furthermore, the magnitudes of effect of these variables were robust: per unit decrease in BMI, a patient could
Table 2 Laboratory data and MM staging
Table 3 Fracture characteristics
Characteristics Laboratory data CRP (mg/dL) ESR (mm/h) b2-MG (mg/L) Serum albumin (g/dL) Hypoalbumenic Serum M-protein (g/dL) Serum light chains (g/L) Serum calcium (mg/dL) Corrected serum calcium (mg/dL) Hypercalcemic Serum creatinine (mg/dL) Serum hemoglobin (g/dL) Anemic* Marrow plasma cell (%) ISS Stage 1 Stage 2 Stage 3 DSS Stage 1 Stage 2 Stage 3
Statistics 1.2262.14 66651 4.5763.06 3.6960.67 20 (40) 2.4561.86 0.88 9.6661.29 9.7161.85 17 (34) 1.0060.42 10.9762.47 35 (70) 45628 14 (28) 14 (28) 22 (44) 0 (0) 32 (64) 18 (36)
b2-MG, b2-microglobulin; CRP, C-reactive protein; DSS, DurieSalmon Staging System; ESR, erythrocyte sedimentation rate; ISS, International Staging System; M-protein, myeloma protein. Note: Values are presented as mean6standard deviation or n (%). * Gender adjusted.
Characteristics
Statistics
Number of fractures Number of patients with 1 fracture 2 fractures 3–5 fractures O5 fractures Percent vertebral body height loss* Anterior Middle Posterior Genant grade 1 2 3 Fracture morphology Wedge Biconcave Crush Location Cervical Thoracic Lumbar
124 28 9 9 4
(56) (18) (18) (8)
30621 37621 16618 25 (20) 39 (32) 60 (48) 40 (32) 68 (55) 16 (13) 1 (1) 83 (67) 40 (32)
Note: Values are presented as mean6standard deviation or n (%). * Height loss is relative to average of adjacent levels (eg, average of T4 and T6 heights for a T5 fracture). Anterior measurement is height of vertebral body at anterior edge on sagittal midline T1-weighted magnetic resonance imaging slice at midline. The middle and posterior measurements are in the middle and posterior aspects of the vertebral body, respectively.
J.A. Miller et al. / The Spine Journal
-
(2015)
-
5
Fig. 2. Vertebral fractures by location. Distribution of 124 observed fractures plotted with respect to vertebral bodies.
expect 0.1 additional fractures; per unit (grams per deciliter) decrease in albumin, a patient could expect 0.3 additional fractures; per unit (grams per deciliter) increase in M-protein, a patient could expect 0.1 additional fractures; per unit (grams per liter) increase in light chains, a patient could expect 0.2 additional fractures; and finally, per unit (milligrams per deciliter) increase in creatinine, a patient could expect nearly one additional fracture. Similarly, simple logistic regression was performed to identify predictors of vertebral body height loss. Four independent variables were associated (p!.2) with an increased degree of height loss: a history of smoking, increased CRP, increased b2-MG, and increased creatinine. After multivariable regression, increased b2-MG (p5.04) and increased creatinine (p5.03) remained significantly and independently predictive of greater vertebral body height loss. Per unit increase in b2-MG (milligrams per liter) and creatinine (milligrams per deciliter), a patient could expect 4% and 58% additional height loss of the fractured vertebra, respectively.
Discussion Fracture incidence, characteristics, and radiology Previous studies have characterized the incidence, associated symptoms, and marrow characteristics of vertebral fractures in MM. Melton et al. [21] conducted a population-based study of 165 patients with MM; they found that 134 (81%) of these patients experienced 225 fractures (118 vertebral). This study established a reliable incidence of fractures in an MM population. Also in the present study, we found that vertebral fractures occur at first presentation with MM in 76% of patients. Sparse radiologic data have been reported for MMassociated vertebral fractures. Lecouvet et al. [3] reported the appearance and distribution of vertebral fractures in 37 patients with newly diagnosed MM. Among 224 vertebral fractures, 149 appeared benign and 75 appeared malignant based on MRI findings at any time in the clinical course (initial presentation or follow-up). The majority of these fractures were in the thoracic spine, similar to what was observed in the
Fig. 3. Distribution of dependent variables. (Top) Distribution of the number of patients sustaining single or multiple vertebral fractures. (Bottom) Distribution of the number of patients sustaining various vertebral body height losses.
6
J.A. Miller et al. / The Spine Journal
Table 4 Multivariable regression analysis Multivariable regression No. of presenting fractures* Independent variable
Percent height lossy
b Coefficient p Value Odds ratio p Value
Age at fracture (y) Gender (F) History of smoking BMI L0.103 Hypertension CAD CHF Diabetes CKD Stage $3 Osteopenia/porosis Hx other cancer CRP (mg/dL) ESR (mm/hr) b2-MG (mg/L) 0.351 Serum albumin (g/dL) L0.307 Serum M-protein (g/dL) 0.132 Serum light chains (g/L) 0.173 Serum calcium (mg/dL) Serum creatinine (mg/dL) 0.925 Serum hemoglobin (g/dL) Marrow plasma cell (%) Intl. MM stage [3] DSS [3] 0.347
1.054
.50
!.01
-
(2015)
-
height loss (37%) occurred in the middle of the vertebrae. In the osteoporotic population, it appears that greater anterior height loss is more common, likely as a result of axial loading and kyphosis of the thoracic spine [18]. Although axial loading and spinal curvature are certainly factors in the development of myeloma-associated fractures, the greater prevalence of biconcave fractures may relate to local disease in the marrow promoting interluekin-6–mediated osteolysis and central collapse of the vertebrae [1]. Risk factors for fracture burden and severity
0.930
.20
.94 .02 .01 !.01
1.036
.04
!.01
1.581
.03
.18
BMI, body mass index; Bx, biopsy; CAD, coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; CRP, C-reactive protein; DSS, Durie-Salmon Staging System; ESR, erythrocyte sedimentation rate; F, female; Hx, history; Intl., international; b2-MG, b2microglobulin; MM, multiple myeloma; M-protein, myeloma protein. Note: Bold values are statistically significant (p!.05) * Dependent variable is number of vertebral fractures at presentation. y Dependent variable is greatest percent height loss of the fractured vertebra.
present study. However, this study included only patients with DSS Stage 3 disease and before 1997, before the widespread use of bisphosphonates and modern chemotherapy for the prevention of subsequent fractures [22,23]. In another study, Lecouvet et al. [4] reported the correlation between MRI bone marrow patterns with vertebral fracture risk. They found that four basic marrow patterns existed (A–D, from normal marrow to diffuse marrow involvement) and that patients with more benign-appearing marrow experienced longer fracture-free survival. The relative risks for subsequent fractures with patterns C/D with respect to A/B were between 6.2 and 11.0. More than 90% of patients with patterns C/D experienced fractures within 3 months of MM diagnosis, whereas patients with benign patterns (A/B) experienced fractures later, at a median of 12 months. In the present study, the vast majority (76%) of patients presented with fractures without a previous MM diagnosis. Of those with a previous MM diagnosis, the median time to fracture was 4 months (Fig. 1). In the present series, the majority (55%) of fractures were biconcave in nature, and as such, the greatest mean
In a retrospective population-based review of 145 patients diagnosed with MM from 1945 to 2001, Melton et al. [21] reported that the average MM patient was anemic (11.4 g/dL) and normocalcemic (9.4 g/dL), with slight renal insufficiency (creatinine, 1.2 mg/dL) and a 40% bone marrow plasma cell burden. These values are similar to those observed in the present study. Using these data, Melton et al. [21] performed a multivariable analysis and observed that serum calcium, corticosteroid use, and chemotherapy were risk factors for subsequent fractures of any location. However, no laboratory values other than serum calcium were significantly associated with fracture risk. Lecouvet et al. [4] performed a similar multivariable analysis attempting to identify predictors of fracture-free survival in 37 newly diagnosed DSS Stage 3 MM patients. It was observed that increased b2-MG and cumulative steroid intake predicted first and second fracture-free survival, respectively. The present study identified risk factors that expand on the reports of Melton et al. [21] and Lecouvet et al. [4]. Two outcome variables were used: the number of presenting vertebral fractures and greatest height loss of the fractured vertebral body (anterior, middle, or posterior). Because 76% of patients presented with fracture without a previous diagnosis of MM, the present study did not investigate risk factors for first fracture. Decreased BMI and albumin and increased M-protein, light chains, and creatinine were found to be statistically significant predictors of multiple fractures at presentation after adjusting for confounders using a Poisson model. The magnitude of effect associated with these variables was substantial: for example, a patient with a BMI of 23 may expect one additional fracture compared with a patient with a BMI of 33. The wasting of malignancy or the greater prevalence of osteoporosis in patients with low BMI may explain this finding. As such, BMI and low albumin seem to be crude measures of advanced disease, and, therefore, potentially greater number of fractures. In addition, increases in myeloma-specific laboratory values (M-protein and light chains) and renal insufficiency carried a substantial risk for multiple fractures at presentation. These results suggest that patients with laboratory signs of more advanced disease may necessitate earlier or more frequent imaging. The present study also investigated independent predictors of more severe fractures at presentation. Increased b2-MG and
J.A. Miller et al. / The Spine Journal
creatinine were statistically significant independent predictors after adjusting for confounders. This likewise suggests that laboratory markers of MM are acceptable surrogates for identifying severe fractures. The clinician/surgeon may suspect MM in patients with back pain, decreased BMI and albumin, and increased creatinine, thereby avoiding excess or unnecessary imaging and laboratory tests in all patients with MM. Given the results of the present study and data reported in the literature, it appears that diagnosing MM before first fracture is exceedingly challenging, as approximately 75% of patients present with fracture at diagnosis. Furthermore, myeloma is a heterogeneous disease with different rates of progression, and routine screening for MM in healthy individuals is not performed. As such, MM patients generally present after onset of constitutional symptoms or laboratory anomalies (anemia, renal insufficiency, and so forth), and a high clinical suspicion for the disease is required. Identifying MM earlier in the disease course may prevent these fractures, which require an unknown duration of osteolysis before occurring. Given the identified risk factors in the present study, it appears that a number of accessible and routine measures are independently associated with fracture number and severity, including BMI, serum albumin, and serum creatinine. Although the specificity of these three variables is quite low, it is possible to develop nomograms from such combined data to determine the pretest probability of a myeloma-associated fracture. With these proposed nomograms and a high suspicion for myeloma, primarycare providers may proceed to sensitive and specific assays, such as b2-MG, M-protein, or light chains, for early identification of patients with myeloma. Alternatively, simple radiographs of the spine may be performed for fracture screening in patients with back pain demonstrating low BMI, low albumin, and elevated creatinine. Although initial vertebral fractures may not be preventable in all MM patients, earlier referral to spine clinics may lessen the morbidity associated with these fractures. Concomitantly, earlier use of bisphosphonates or RANK ligand inhibitors may serve to prevent or delay the progression of these fractures. Moreover, with available tools such as the Spinal Instability Neoplastic Score, clinicians are identifying patients at greatest risk for impending vertebral fracture [24]. In the myeloma population, a combined approach using dual-energy X-ray absorptiometry scores, Spinal Instability Neoplastic Score, and patient BMI and laboratory data is certainly feasible and practical, as these data are routinely collected. With this knowledge, prophylactic stabilization with established techniques such as kyphoplasty may alleviate bony morbidity before clinical deterioration. Limitations The present study had several limitations. Because of the retrospective nature of the investigation, data reported by clinicians were not standardized and subject to bias. Furthermore, this study was not a cohort study, and, therefore, did not
-
(2015)
-
7
follow a population of MM patients until first vertebral fracture. However, our data suggest that patients with MM are rarely diagnosed before development of a first vertebral fracture. It is also possible that there were patients who did not receive regular bone surveys; therefore, fractures were not identified in others who died before fracture detection. As such, a subset of undiagnosed fractures may not be characterized in this study. Last, although many variables were controlled for in the multivariable logistic regression, there may be unmeasured variables that confound the results. However, this study included variables related to demographics, comorbidities, laboratory values, and staging in an effort to identify any contributing confounding variables. The independent predictors identified in this study were not only statistically significant but biologically and clinically relevant in their relation to outcome variables.
Conclusions Vertebral fractures in MM are significantly detrimental to patient quality of life in a disease with a 45% five-year survival. Despite the burden of these fractures, they have been poorly characterized radiologically, and few independent risk factors for fracture burden have been established. In the present study, vertebral fractures associated with MM were generally severe, biconcave in morphology, and in the thoracic spine. Furthermore, in almost all patients, vertebral fracture was identified at the time of initial MM diagnosis. Given this, it is imperative to identify fractures early in their natural history such that medical and surgical interventions may be conducted. To aid in this objective, a number of independent risk factors for fracture burden and severity at presentation were identified, including lower BMI and albumin and higher b2-MG, M-protein, light chains, and creatinine. Based on these findings, it may be feasible to use nomograms of BMI, albumin, and creatinine to determine which patients in the primary-care setting are at greatest risk for MM. Subsequent laboratory testing with specific assays such as b2-MG may then diagnose myeloma and associated fractures earlier in the disease course. Future studies observing the progression of these fractures over clinical follow-up are warranted to characterize their rate of change and to establish evidence-based guidelines for spine surveillance in the myeloma population. References [1] Angtuaco EJC, Fassas ABT, Walker R, Sethi R, Barlogie B. Multiple myeloma: clinical review and diagnostic imaging. Radiology 2004;231:11–23. [2] Palumbo A, Anderson K. Multiple myeloma. N Engl J Med 2011;364:1046–60. [3] Lecouvet FE, Vande Berg BC, Maldague BE, Michaux L, Laterre E, Michaux JL, et al. Vertebral compression fractures in multiple myeloma. Part I. Distribution and appearance at MR imaging. Radiology 1997;204:195–9.
8
J.A. Miller et al. / The Spine Journal
[4] Lecouvet FE, Malghem J, Michaux L, Michaux JL, Lehmann F, Maldague BE, et al. Vertebral compression fractures in multiple myeloma. Part II. Assessment of fracture risk with MR imaging of spinal bone marrow. Radiology 1997;204:201–5. [5] Erdem E, Akdol S, Amole A, Fryar K, Eberle RW. Radiofrequencytargeted vertebral augmentation for the treatment of vertebral compression fractures as a result of multiple myeloma. Spine 2013;38:1275–81. [6] Dudeney S, Lieberman IH, Reinhardt M-K, Hussein M. Kyphoplasty in the treatment of osteolytic vertebral compression fractures as a result of multiple myeloma. J Clin Oncol 2002;20:2382–7. [7] Mendoza TR, Koyyalagunta D, Burton AW, Thomas SK, Phan MH, Giralt SA, et al. Changes in pain and other symptoms in patients with painful multiple myeloma-related vertebral fracture treated with kyphoplasty or vertebroplasty. J Pain 2012;13:564–70. [8] Julka A, Tolhurst SR, Srinivasan RC, Graziano GP. Functional outcomes and height restoration for patients with multiple myeloma-related osteolytic vertebral compression fractures treated with kyphoplasty. J. Spinal Disord Tech 2014;27:342–6. [9] Pflugmacher R, Kandziora F, Schroeder RJ, Melcher I, Haas NP, Klostermann CK. Percutaneous balloon kyphoplasty in the treatment of pathological vertebral body fracture and deformity in multiple myeloma: a one-year follow-up. Acta Radiol 2006;47:369–76. [10] McDonald RJ, Trout AT, Gray LA, Dispenzieri A, Thielen KR, Kallmes DF. Vertebroplasty in multiple myeloma: outcomes in a large patient series. AJNR Am J Neuroradiol 2008;29:642–8. [11] Berenson J, Pflugmacher R, Jarzem P, Zonder J, Sjchechtman K, Tillman JB, et al. Balloon kyphoplasty versus non-surgical fracture management for treatment of painful vertebral body compression fractures in patients with cancer: a multicentre, randomised controlled trial. Lancet Oncol 2011;12:225–35. [12] Griffith JF, Adams JE, Genant HK. Diagnosis and classification of vertebral fracture. In: Rosen CJ, ed. Primer on the metabolic bone diseases and disorders of mineral metabolism. John Wiley & Sons, Inc., 2013: 317–35. Available at: http://onlinelibrary.wiley.com/doi/10.1002/ 9781118453926.ch37/summary. Accessed September 18, 2014. [13] Fechtner K, Hillengass J, Delorme S, Heiss C, Neben K, Goldschmidt H, et al. Staging monoclonal plasma cell disease:
[14]
[15] [16]
[17]
[18]
[19]
[20] [21]
[22] [23]
[24]
-
(2015)
-
comparison of the Durie-Salmon and the Durie-Salmon PLUS staging systems. Radiology 2010;257:195–204. Hari PN, Zhang M-J, Roy V, Perez WS, Bashey A, To LB, et al. Is the International Staging System superior to the Durie-Salmon staging system? A comparison in multiple myeloma patients undergoing autologous transplant. Leukemia 2009;23:1528–34. Anderson KC. Multiple myeloma: a clinical overview. Oncology (Williston Park) 2011;25(Suppl 2):3–9. Adams JE, Lenchik L, Roux C, Genant HK. Vertebral fracture initiative: radiologic assessment of vertebral fracture. Available at: http:// www.iofbonehealth.org/sites/default/files/PDFs/Vertebral%20Fractur e%20Initiative/IOF_VFI8Part_II-Manuscript.pdf. Accessed September 18, 2014. Grados F, Fechtenbaum J, Flipon E, Kolta S, Roux C, Fardellone P. Radiographic methods for evaluating osteoporotic vertebral fractures. Joint Bone Spine 2009;76:241–7. Jarraya M, Hayashi D, Griffith JF, Guermazi A, Genant HK. Identification of vertebral fractures. In: Guglielmi G, ed. Osteoporosis and bone densitometry measurements. Medical radiology. Berlin, Germany: Springer, 2013:41–55. Available at: http://link.springer.com/ chapter/10.1007/174_2012_701. Accessed September 18, 2014. Vaccaro AR, Oner C, Kepler CK, Dvorak M, Schnake K, Bellabarba C, et al. AOSpine thoracolumbar spine injury classification system: fracture description, neurological status, and key modifiers. Spine 2013;38:2028–37. JMP statistical software. Cary, NC: SAS Institute, Inc., 1989. Melton LJ III, Kyle RA, Achenbach SJ, Oberg AL, Rajkumar SV. Fracture risk with multiple myeloma: a population-based study. J Bone Miner Res 2005;20:487–93. Berenson JR, Lipton A. Bisphosphonates in the treatment of malignant bone disease. Annu Rev Med 1999;50:237–48. Berenson JR, Lipton A. Use of bisphosphonates in patients with metastatic bone disease. Oncology (Williston Park) 1998;12: 1573–9; discussion 1579–1581. Fourney DR, Frangou EM, Rkyen TC, Dipaola CP, Shaffrey CI, Berven SH, et al. Spinal instability neoplastic score: an analysis of reliability and validity from the spine oncology study group. J Clin Oncol 2011;29:3072–7.
