Predictors of new fragility fractures after diagnosis of indolent systemic mastocytosis Eveline van der Veer, PhD,a Suzanne Arends, PhD,b Sjoukje van der Hoek, BSc,c Joris B. Versluijs, BSc,a Jan G. R. de Monchy, MD, PhD,c Joanna N. G. Oude Elberink, MD, PhD,c,d and Jasper J. van Doormaal, MD, PhDc Groningen, The Netherlands Background: Fragility fractures (FFxs) and osteoporosis occur frequently in patients with indolent systemic mastocytosis (ISM), even before 50 years of age. Objective: We sought to develop a prediction model to identify individual patients with ISM at risk of new FFxs. Methods: Data on lifetime fractures and trauma circumstances were collected from vertebral morphometry, patients’ records, and questionnaires. Clinical, lifestyle, and bone characteristics were measured. Patients receiving treatment for osteoporosis before ISM diagnosis or with missing bone data were excluded from FFx risk assessment. Results: In total, 389 lifetime fractures occurred in 127 of the 221 patients with ISM (age, 19-77 years), including 90 patients with 264 FFxs. Median follow-up after diagnosis was 5.4 years (range, 0.4-15.3 years), with 5- and 10-year FFx risks of 23% 6 3% and 31% 6 4%, respectively. Male sex, high levels of bone resorption (serum type I collagen C-telopeptide), low hip bone mineral density, absence of urticaria pigmentosa, and alcohol intake at the time of ISM diagnosis were independent predictors of future FFxs. The MastFx score, a prediction model using these 5 characteristics, showed good accuracy (area under the curve, 0.80) to determine the risk of new FFxs. QFracture, a validated risk scoring tool for persons aged 30 to 99 years, was not useful in patients with ISM. Conclusion: The MastFx score distinguishes patients with ISM at high, intermediate, and low risk of new FFxs. The included characteristics sex, serum type I collagen C-telopeptide, hip bone mineral density, urticaria pigmentosa, and alcohol intake are easy to collect in clinical practice. The high occurrence of FFxs in patients with ISM underlines the importance of optimizing bone quality and early start of therapeutic prevention in patients at risk. (J Allergy Clin Immunol 2014;134:1413-21.) Key words: Systemic mastocytosis, fragility fractures, fracture risk scoring tool, bone turnover markers, bone mineral density, alcohol, sex, urticaria pigmentosa, serum type I collagen C-telopeptide
From aLaboratory Medicine, bRheumatology and Clinical Immunology, cAllergy, and d the GRIAC Research Institute, University of Groningen, University Medical Center Groningen. Disclosure of potential conflict of interest: The authors declare that they have no relevant conflicts of interest. Received for publication September 13, 2013; revised April 3, 2014; accepted for publication May 8, 2014. Available online June 27, 2014. Corresponding author: Eveline van der Veer, PhD, Department of Laboratory Medicine, HPC-EA40, University Medical Center Groningen, PO Box 30.001, 9700 RB Groningen, The Netherlands. E-mail: [email protected]. 0091-6749/$36.00 Ó 2014 American Academy of Allergy, Asthma & Immunology http://dx.doi.org/10.1016/j.jaci.2014.05.003
Abbreviations used AUC: Area under the curve BMD: Bone mineral density BMI: Body mass index BTM: Bone turnover marker CVinter: Interassay coefficient of variation FFx: Fragility fracture FRAX: Fracture Risk Assessment Tool HR: Hazard ratio ISM: Indolent systemic mastocytosis MH: Methylhistamine MIMA: Methylimidazole acetic acid ROC: Receiver operating characteristic sCTX: Serum type I collagen C-telopeptide UP: Urticaria pigmentosa
Indolent systemic mastocytosis (ISM) is defined by the presence of abnormal mast cells outside the skin, principally in the bone marrow.1 Mast cells produce and release a large number of different mediators, such as histamine, prostaglandin D2, platelet-activating factor, proinflammatory cytokines, leukotrienes C4 and D4, chemokines, and tryptase.2 Symptoms and signs are caused by release of these mast cell mediators, local accumulation of mast cells, or both. Manifestations vary strongly in nature and severity and can include skin and gastrointestinal symptoms, anaphylaxis, musculoskeletal pain, and neuropsychiatric symptoms. Fragility fractures (FFxs), osteoporosis, or both occur in approximately half of the patients with ISM.3-9 Notably, the prevalence of FFxs was high in men, even in those less than 50 years of age.4 Mast cell mediators are branded to have deleterious effects on bone metabolism.10,11 In addition, activated mast cells release the proinflammatory cytokines IL-1, IL-6, and TNF-a,12 which are important regulators of bone resorption.10 Some patients with ISM exhibit increased serum IL-6 levels, and a correlation with disease severity was found in a group of patients with different types of mastocytosis.13,14 Furthermore, alcohol triggers mast cell degranulation.15-17 Its intake can be associated with a number of adverse reactions in patients with ISM. Apart from toxicological effects, intolerance occurs and might manifest clinically as flushing and, in rare cases, as generalized urticaria and anaphylactic reactions after ingestion. In the general population multiple tools have been developed to assess a patient’s fracture risk based on models that integrate age and sex with additional risk factors (eg, low body mass index [BMI], previous FFxs, underlying diseases, use of corticosteroids, smoking, and excessive alcohol intake), with or without the use of bone mineral density (BMD) at the femoral neck. Recently, the National Institute of Clinical Excellence’s guidelines 1413
1414 VAN DER VEER ET AL
recommended assessing bone health by using one of the 2 validated scoring systems, namely the Fracture Risk Assessment Tool (FRAX) and QFracture.18 However, these fracture risk scoring systems do not take into account specific risk factors associated with FFxs in the younger population. FRAX is limited to persons older than 40 years. Even though QFracture includes persons between 30 and 40 years of age, the score is less reliable because the number of fractures in the data set is small for this age group. In addition, the prevalence of multiple fractures in patients with ISM3,4 is much higher than the prevalence found in European population studies.19-23 Therefore the aim of the present study was to investigate the predictive value of certain patient characteristics at the time of diagnosis for the occurrence of future FFxs in patients with ISM. By using these characteristics, an easyto-use prediction model, named the MastFx score, was developed and tested to identify patients with ISM at risk of new FFxs.
METHODS Study population Between 1981 and August 2012, ISM was diagnosed in 228 consecutive patients at the referral centre for mastocytosis of the University Medical Center Groningen. The diagnosis was established by fulfilling the criteria of the World Health Organization Classification of Mastocytosis. Patients with aggressive forms of systemic mastocytosis were excluded. Smoldering systemic mastocytosis, a provisional subentity of ISM with detectable B findings (criteria used for classification of systemic mastocytosis) caused by the mast cell disease process, was excluded as well.24 Fracture data were collected from medical records and lateral radiographs of the thoracolumbar spine combined with a questionnaire sent to all patients. FFxs, also known as low-level (or ‘‘low-energy’’) trauma fractures, were defined as fractures that result from mechanical forces that would not ordinarily result in fractures.25 In cases in which it was not possible to discriminate between high- or low-energy trauma circumstances, a telephone call was made to clarify. Fracture data of 7 (3%) of the 228 patients could not be retrieved, and these patients were excluded. BMD of the lumbar spine (anterior-posterior projection at L1-L4) and hip (total proximal femur) were measured with dual energy x-ray absorptiometry (DXA) (Hologic QDR Discovery, Waltham, Mass). BMD T-scores were calculated by using the National Health and Nutrition Examination Survey reference database. Osteoporosis was defined according to World Health Organization guidelines as a BMD T-score of 22.5 or less, and osteopenia was defined as a BMD T-score of between 21.0 and 22.5. In addition, data regarding osteosclerosis, the presence of urticaria pigmentosa (UP), and anaphylaxis were collected (definitions, details, or both were published previously4). For the predictor analysis, 40 of the 221 patients were excluded because of treatment for osteoporosis (eg, bisphosphonates or parathyroid hormone) before ISM diagnosis (n 5 28), sex change (n 5 1), fracture or surgery within the last 6 months before diagnosis (n 5 2), or missing BMD and bone turnover marker (BTM) data (n 5 9, Fig 1). The Medical Ethical Review Board of the University Medical Center Groningen declared that the study was performed in accordance with regulations of the review board for publication of subjects’ data.
Biochemical markers BTM was studied by means of assessment of the formation markers bonespecific alkaline phosphatase, procollagen type 1 N-terminal peptide, and osteocalcin and the resorption marker serum type I collagen C-telopeptide (sCTX).26 Bone-specific alkaline phosphatase levels were measured by using ELISA (Metra Biosystems, Mountain View, Calif; interassay coefficient of variation [CVinter], 5.5%), procollagen type 1 N-terminal peptide levels were measured by using an RIA (Orion Diagnostica, Espoo, Finland; CVinter, 9.0%), osteocalcin levels were measured by using an immunoradiometric assay (BioSource Europe S.A, Nivelles, Belgium; CVinter, 9.4%), and sCTX
J ALLERGY CLIN IMMUNOL DECEMBER 2014
levels were measured by using an electrochemiluminescence immunoassay (Elecsys 2010; Roche, Mannheim, Germany; CVinter, 10.8%). Serum samples were stored within 1 hour after venipuncture at 2208C until analysis. BTM Z-scores, the number of SDs from the normal mean corrected for age and sex, were calculated by using a Dutch reference group (200 men and 350 women) checked for serum 25-hydroxyvitamin D levels of greater than 50 nmol/L, as well as for the absence of osteoporosis (BMD T-score, >22.5) after 50 years of age. Z-scores were calculated as follows: (BTM value of individual patient 2 Mean BTM value of matched 10-year cohort of healthy reference group)/SD of this matched reference cohort. Tryptase levels were determined by using the B12 assay (Pharmacia UniCAP Tryptase; Thermo Fisher Scientific/Pharmacia and Upjohn, Uppsala, Sweden; CVinter, 5.8%).27 Reference values for healthy subjects are from Pharmacia and Upjohn, showing a geometric mean level of 3.8 mg/L and an upper 95th percentile of 11.4 mg/L. Urine samples were collected after an overnight fast, discarding the first voiding after wakening, to measure methylhistamine (MH) and methylimidazole acetic acid (MIMA) values. For 24 hours before urine collection, patients were asked to refrain from histamine-rich foods and drinks, such as sauerkraut, canned fish, yogurt, and wine. The MH value was determined by using an isotope-dilution mass fragmentographic method.28 MIMA was determined as described previously,29 with some modifications using isotope dilution mass fragmentography. For MH and MIMA excretion, mean 6 SD values of an apparently healthy population were 101 6 33 mmol/mol (range, 50-154 mmol/mol; CVinter, 6.8%) and 1.3 6 0.3 mmol/mol (range, 0.9-1.9 mmol/mol; CVinter, 4.2%) creatinine, respectively.30
Statistical analysis Statistical analysis was performed with IBM SPSS Statistics 20 software (SPSS, Chicago, Ill). Results were expressed as means 6 SDs or medians (ranges) for parametric and nonparametric data, respectively. Predictor analysis for the occurrence of FFxs was performed by using univariate and multivariate Cox regression with backward Wald inclusion of variables with a P value of .20 or less in univariate analysis. The P value for stepwise removal was .10, and P values of less than .05 were considered statistically significant. For continuous variables with significant outcomes in univariate analysis, relevant cutoff points were determined according to clinical applicability and checked by using receiver operating characteristic (ROC) curves. These categorized variables were used in multivariate analysis. The final prediction model, named the MastFx score, was obtained by rounding the regression coefficients of the multivariate model. ROC analysis was used to determine the performance of this prediction model (area under the curve [AUC] < 0.70 was interpreted as poor, 0.70 < AUC < 0.80 was interpreted as moderate, 0.80 < AUC < 0.90 was interpreted as good, and AUC > 0.90 was interpreted as high accuracy). Finally, the MastFx scores were compared with risk scores calculated with one of the National Institute of Clinical Excellence–recommended validated scoring tools, the QFracture-2012 algorithm (version 1.0; ClinRisk Ltd, Leeds, United Kingdom). For patients with ISM aged less than 30 years (n 5 22), the QFracture risk was calculated with an age of 30 years.
RESULTS FFxs Information about lifetime fractures was retrieved for 221 (97%) patients. In total, 389 lifetime fractures occurred in 127 patients. Of these fractures, 125 were high-energy trauma fractures, and 264 were FFxs. Before the ISM diagnosis was made, 54 patients reported 139 FFxs. After the ISM diagnosis, 56 patients reported 125 new FFxs during a median follow-up of 5.4 years (range, 0.4-15.3 years; Fig 2, A), with a 5- and 10-year fracture-free survival of 77% 6 3% and 69% 6 4%, respectively. Forty patients were excluded from the predictor analysis to avoid bias of factors influencing fracture risk, BMD, and/or BTM (Fig 1). In the remaining 181 patients with ISM, 27 reported 40
VAN DER VEER ET AL 1415
J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 6
FIG 1. Flowchart of handling of fracture data from patients with ISM.
FFxs before ISM diagnosis, and 43 reported 88 FFxs after diagnosis (Fig 2, B). Median follow-up (5.3 years; range, 0.4-15.3 years) and 5- and 10-year fracture-free survival (78% 6 4% and 71% 6 4%) were comparable with those of the total group.
Patient characteristics Characteristics at the time of diagnosis of the 181 patients with ISM included in the predictor analysis are presented in Table I. Mean age was 46 6 13 years, and 59% of the patients were female. Only 3 women had a BMI of less than 19 kg/m2. Another 14 women weighed less than 60 kg, although they had normal BMI. Current smoking or smoking stopped within the last 10 years was reported by 64% of the patients, with differences in smoking habits between male and female subjects (58% vs 38%, P 5.04). Seventy-six percent reported social alcohol intake, but only 8% used 3 or more units per day. Thirty-eight patients with ISM reported mild-to-moderate adverse reactions after past intake of alcohol. Median duration between first symptoms and diagnosis was 7.5 years (range, 0.1-58 years). UP was present in 76% of the patients, 43% experienced anaphylactic reactions, and 35% used at least 1 antimediator drug.
Osteoporosis of the lumbar spine and hip was detected in 14% and 0.6%, respectively, and 15% of the patients had a previous FFx. The 43 patients with new FFxs during follow-up had normal BMD (n 5 12 [28%]), osteopenia (n 5 20 [47%]), or osteoporosis (n 5 11 [26%]) at the time of diagnosis.
Predictors Patients with new FFxs after ISM diagnosis were older; were more often male; had more frequent anaphylactic reactions; had UP less often; had higher MIMA, osteocalcin, and sCTX levels; had lower hip BMD scores; and reported more frequent alcohol intake at the moment of diagnosis (Table II). For clinical applicability, relevant cutoff points were determined (Table II) and checked by using ROC curves (data not shown). Multivariate Cox regression analysis showed that male sex (hazard ratio [HR], 2.04), a sCTX Z-score of 11.0 or greater (HR, 2.63), a hip BMD T-score of 21.0 or less (HR, 2.19), absence of UP (HR, 2.05), and alcohol intake (HR, 3.45) at the time of ISM diagnosis were independent predictors of future FFxs (Table III). The fracture-free survival curves for each of the predictors are shown in Fig 3. Further investigation revealed that only 33% of
1416 VAN DER VEER ET AL
J ALLERGY CLIN IMMUNOL DECEMBER 2014
FIG 2. Lifetime FFxs in patients before and/or after the diagnosis of ISM. A, All patients with ISM (n 5 221). B, Patients with ISM used for the predictor analysis (n 5 181).
the patients without UP were less than 50 years of age compared with 66% in the UP group (mean age, 53.8 6 11.6 vs 43.9 6 12.1 years; P < .001). Forty-seven percent of the patients without UP were male compared with 40% in the UP group (P 5 .39).
MastFx score The fracture prediction model, including male sex, high sCTX levels, low hip BMD, absence of UP, and alcohol intake at the time of diagnosis, was used to calculate the MastFx score for each patient by using the regression coefficients of the multivariate model. The accuracy of this model to predict which patients will have fragile fractures was good, with an AUC of 0.80 (95% CI, 0.73-0.88). Comparable results were found if, for clinical applicability, the regression coefficients were rounded to 1 point (AUC, 0.80; 95% CI, 0.72-0.87). Fracture-free survival curves for all patients receiving a MastFx score of 0 to 5 are shown in Fig 4. Patients with a score of 0 (n 5 22) or 1 (n 5 53) had a 5- and 10-year fracture-free survival of 93%; a score of 2 points (n 5 58) resulted in 5- and 10year fracture-free survivals of 78% and 73%, respectively; and a score of 3 points (n 5 35) resulted in 5- and 10-year fracturefree survivals of 65% and 46%, respectively. A small group with a MastFx score of 4 (n 5 12) or 5 (n 5 1) had a 5-year fracture-free survival of 14%. The QFracture tool was used to compare the MastFx score with an already existing and validated scoring system for the risk of FFxs in the general population. The median 5- and 10 year risks of 181 patients with ISM for osteoporotic fracture (hip, spine, wrist, or shoulder) were 0.5% (0.1% to 5.4%) and 1.2% (0.2% to 13.6%)
and the risks for hip fracture were only 0.0% (0.0 to 2.9%) and 0.1% (0.0 to 8.7%), respectively. The risk scores had poor accuracy, with AUC values of 0.60 (95% CI, 0.50-0.71) for major osteoporotic fractures and 0.66 (95% CI, 0.56-0.75) for hip fractures. However, QFracture is not recommended for patients who have experienced an osteoporotic fracture. Excluding 27 patients with ISM with a history of FFxs, the median 5- and 10-year risks of osteoporotic fracture (hip, spine, wrist, or shoulder) were 0.5% (0.1% to 5.4%) and 1.2% (0.3% to 13.6%), respectively, and the median 5- and 10-year risks of hip fracture were only 0.0% (0.0% to 1.9%) and 0.1% (0.0% to 5.5%), respectively. The accuracy for the QFracture score was poor, and only the prediction of hip fracture reached significance (data not shown).
DISCUSSION Patients with ISM have a higher risk of fractures and osteoporosis compared with the general population. In the present study only a quarter of the patients who experienced new fractures after diagnosis had osteoporosis at the time of diagnosis. Thus BMD alone does not have sufficient power to predict the risk of new fractures and to make the clinical decision of whom to treat preventively in this patient group. A prediction model was developed using patient characteristics at the time of diagnosis to be able to make a better estimate of the risk of fractures. The characteristics included in this model, named the MastFx score, are male sex, high sCTX levels, low hip BMD, absence of UP, and alcohol intake, and these 5 parameters can easily be obtained in daily clinical practice. The MastFx score shows good accuracy
VAN DER VEER ET AL 1417
J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 6
TABLE I. Characteristics at the time of diagnosis of 181 patients with ISM Patient characteristics Age (y) Male sex Length (cm) Weight (kg) BMI (kg/m2) Smoking Alcohol use Mastocytosis characteristics Disease duration (y) UP Anaphylactic shock Tryptase (mg/L) MH (mmol/mol creatinine) MIMA (mmol/mol creatinine) Bone characteristics Patients with FFxs No. of FFxs Osteosclerosis Osteocalcin (mg/L) Osteocalcin Z-score (SD) BALP (U/L) BALP Z-score (SD) PINP (mg/L) PINP Z-score (SD) sCTx (pg/mL) sCTX Z-score (SD) LS BMD (g/cm2) LS BMD T-score (SD) Hip BMD (g/cm2) Hip BMD T-score (SD) Comorbidity Patients with comorbidity Diabetes mellitus type 2 Hypothyroidism Hyperthyroidism Hypogonadism Hyperparathyroidism Hypophosphatemia Pulmonary disease Epilepsy Malignancy Rheumatic diseases Crohn disease Other diseases Alcohol abuse Drug abuse Current use of antimediator drug Patients using antimediators H1- and/or H2-antihistamines Cromoglicic acid Oral corticosteroids Inhaled corticosteroids
46 6 13 74 (41%) 174 6 10 79 6 15 26.3 6 4.4 82 (46%) 135 (75%) 7.5 (0.1 to 58) 138 (76%) 77 (43%) 27 (4.1 to 296) 257 (70 to 2554) 3.2 (0.5 to 21.6) 27 (15%) 40 10 (6%) 12.3 (3.6 to 37.2) 20.12 (22.66 to 4.15) 21.5 (4.6 to 62.1) 1.22 (22.09 to 8.55) 39.9 (15.9 to 135) 20.04 (22.10 to 6.62) 195 (10 to 797) 20.42 (22.03 to 6.35) 0.96 6 0.15 20.97 6 1.39 0.93 6 0.13 20.35 6 0.99
_50 y Age >
77 (42%)
Weight 20 mg/L MH >167 mmol/mol creatinine MIMA >1.9 mmol/mol creatinine
(8%) (1.6%) (19%) (30%) (16%) (21%)
122 (67%) 141 (78%) 148 (82%)
Osteocalcin Z-score >12.0 SD
10 (6%)
BALP Z-score >12.0 SD
57 (32%)
PINP Z-score >12.0 SD
12 (7%)
sCTX Z-score >12.0 SD
14 (8%)
Osteoporosis LS
25 (14%)
Osteoporosis hip
1 (0.6%)
48 (27%) 5 4 3 2 1 1 19 1 3 4 2 12 3 2 63 (35%) 55 (31%) 11 (6%) 4 (2.2%) 7 (3.9%)
Values are presented as means 6 SDs, medians (ranges), or numbers (percentages) unless otherwise indicated. BALP, Bone-specific alkaline phosphatase; LS, lumbar spine; PINP, procollagen type 1 N-terminal peptide.
(AUC, 0.80) to determine the risk of future FFxs in patients with ISM. The predictive accuracy of the MastFx score in our data set is much better than that of QFracture, which is a validated tool for assessing the fracture risk in persons aged between 30 and 99
years without a history of osteoporotic fracture.18 For comparison with the MastFx model, patients with ISM and a history of FFxs were included, but the QFracture accuracy was not altered substantially without them. The QFracture model is based on databases of the United Kingdom population. Because no validated
1418 VAN DER VEER ET AL
J ALLERGY CLIN IMMUNOL DECEMBER 2014
TABLE II. Univariable Cox regression to identify parameters that are related to new FFxs occurring after the ISM diagnosis 95% CI Variable
95% CI
HR
Lower
Upper
P value
Cutoff
HR
Lower
Upper
P value
1.015
1.069
.002
_50 y
1 1.995
1.087
3.660
.026
< _4 mmol/mol creatinine >4 mmol/mol creatinine
1 0.998
0.540
1.844
.995
_1.0 SD
1.823
0.918
3.622
.086
_1.0 SD
1 2.929
1.562
5.494
.001
>21.0 SD < _21.0 SD
1 2.496
1.366
4.563
.003
Age
Y
1.042
Sex
Female Male cm kg kg/m2
1 2.554 1.014 1.011 1.016
Absent Present Alcohol use Absent Present Disease duration y Anaphylactic reaction Absent Present UP Absent Present Tryptase mg/L MH (mmol/mol creatinine)/10 MIMA mmol/mol creatinine
1 1.191 1 4.688 0.982 1 2.033 2.602 1 1.003 1.007 1.084
FFxs
Osteocalcin Z-score
n/Patient Absent Present SD
BALP Z-score PINP Z-score
1.379 0.983 0.99 0.954
4.732 1.046 1.031 1.082
.003 .379 .306 .619
0.649
2.184
.573
1.448 0.946
15.18 1.018
.010 .322
1.094 1.423
3.780 4.759
.025 .002
0.998 0.998 1.002
1.009 1.015 1.174
.263 .110 .044
1.366 1 1.648 1.272
0.991
1.882
.057
0.810 1.010
3.352 1.601
.168 .041
SD SD
0.981 1.166
0.806 0.923
1.194 1.473
.851 .199
sCTX Z-score
SD
1.217
1.026
1.442
.024
LS BMD T-score Hip BMD T-score
SD SD
0.883 0.669
0.715 0.484
1.091 0.924
.250 .015
Osteoporosis*
Absent Present Absent Present Absent Present
1 1.747 1 1.369 1 1.119
Length Weight BMI Smoking
Comorbidity Antimediator use
0.880
3.469
.111
0.723
2.594
.335
0.643
1.946
.691
Values in boldface indicate statistical significance. BALP, Bone-specific alkaline phosphatase; PINP, procollagen type 1 N-terminal peptide; LS, lumbar spine. *Defined as a lumbar spine, hip, or both BMD T score of 22.5 SD or less.
TABLE III. Multivariate Cox regression to identify independent predictors of new FFxs after ISM diagnosis 95% CI
Male sex _11.0 sCTX Z-score > _21.0 Hip BMD T-score < UP absence Alcohol use
HR
Lower
Upper
P value
B
MastFx score
2.043 2.632 2.187 2.047 3.445
1.045 1.278 1.128 1.074 1.016
3.996 5.424 4.240 3.899 11.688
.037 .009 .021 .029 .047
0.715 0.968 0.782 0.716 1.237
1 1 1 1 1
Values in boldface indicate statistical significance. B, Regression coefficient.
risk calculation tools were available for the general Dutch population, we used the country for which the epidemiology of fractures in the general population most closely approximates our country. Because 30% of the patients with ISM were aged less than 40 years, no comparison was made with the FRAX scores. A recent review compared 6 different fracture risk assessment
tools, including QFracture and FRAX, which were externally validated in a population-based setting with proper methodological quality. The authors found that no tool performed consistently better than others (ie, AUC values mostly ranged between 0.6 and 0.8), indicating that all the tools are modest predictors of fractures.31
J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 6
VAN DER VEER ET AL 1419
FIG 3. FFx-free survival curves of the 5 independent predictors: male sex, sCTX Z-score of 11.0 or greater, hip BMD T-score of 21.0 or less, absence of UP, and use of alcohol.
In the general population older age is a well-known risk factor for the development of fractures. In our mastocytosis cohort 24% (54/221) of patients had FFxs at the time of diagnosis, which increased to 41% at the study’s closing date (median follow-up, 5.4 years). Although this indicates the relevance of age, this parameter was not identified as an independent predictor in our MastFx model. Compared with the general population, fractures occurred more frequently in young men with ISM and postmenopausal women with ISM and not in premenopausal women with ISM, as we discussed in our previous study.4 An interesting finding was that the presence of UP protected against fractures.
Further exploration of this finding is needed. In our cohort 49 (27%) of the 181 patients were women less than 50 years of age, 45 of whom had UP. Although only 2 of these patients had new FFxs after ISM diagnosis, no significant interactions were found between age or sex and UP. As expected, high sCTX levels and low hip BMD were associated with FFxs, which is in agreement with earlier publications.32-34 Surprisingly, the mast cell mediators were statistically not selected in the prediction model. Previously, we have shown that higher MH levels were associated with a higher risk of osteoporotic manifestations (osteoporotic fractures and
1420 VAN DER VEER ET AL
FIG 4. FFx-free survival curves per MastFx score (0-5) using a prediction model consisting of 5 patient characteristics.
osteoporosis combined).4 In the present study only MIMA levels had a significant positive association with new osteoporotic fractures in univariate analysis. Perhaps urinary histamine metabolite levels reflect mast cell burden in bone marrow to only a minor extent. Another possibility might be that the deleterious effect of local exposure from the mast cell mediators on the bone is of greater importance than the effect of circulating mediators, as indicated by levels measured in serum and urine. A third explanation might be that the histamine metabolites are lesser risk factors when BTM is taken into account. In different European cohorts the percentage of patients with ISM and FFxs varied from 5% to 37%.3-5,7 These differences can probably be explained by differences in age of the included patients and in the methods used to explore the history and circumstances of lifetime fractures. The 24% history of FFxs in our cohort with a mean age of 48 years is in line with the 21% rate of vertebral fractures in patients with ISM of the same age reported by Italian authors.7 The fracture rate was substantially lower in a younger Spanish cohort.5 In addition, there might be differences in antimediator treatment strategies between the European cohorts. In this study we found no significant difference in future fracture rate between patients with and without antimediator drug use at the time of ISM diagnosis. We were not able to collect data on treatment duration and doses used. Therefore further exploration of the clinical relevance of antimediator drugs on the occurrence of fractures is needed.
J ALLERGY CLIN IMMUNOL DECEMBER 2014
A limitation of this study is the exclusion of 30% of the fractures occurring during the follow-up period by the exclusion of 40 patients, including the 28 who had already been treated for osteoporosis. This method of patient selection might have caused a loss of association with new fractures. Arguments for excluding patients receiving medication for osteoporosis were that these patients were known to be at risk and that measures to reduce the risk of fractures were already taken before the ISM diagnosis was made. In summary, we presented a discriminative and easy-to-use prediction model for the occurrence of FFxs using 5 patient characteristics. Lifetime FFxs were reported by 41% (90/221) of the patients with ISM. Because FFxs can have a devastating effect on the physical and mental condition of the individual patient, appropriate risk assessment is needed. Calculation of the FFx risk should be an important component in the management of patients presenting with ISM. The risk of an individual patient is indicated by the MastFx score, which is calculated as the sum of the following 5 characteristics (1 point each): male sex, high levels of _11.0), low hip BMD (T-score, < _21.0), absence sCTX (Z-score, > of UP, and alcohol intake. On the basis of our prediction model with scores ranging from 0 to 5, patients with ISM with a MastFx score of 2 or greater have a high risk of FFxs. Additional analyses showed that the MastFx score is more useful than an existing fracture risk calculation tool developed for the older general population. Efforts should be made by caretakers to optimize bone quality with adequate vitamin D and calcium intake and lifestyle changes in all patients with ISM. Alcohol cessation in these patients is highly recommended because drinking is a modifiable risk factor for FFxs in patients with ISM. By using the MastFx score, it is possible to distinguish patients with ISM at high fracture risk from those at low fracture risk. These high-risk patients will probably benefit from an early start of therapeutic intervention with drugs, such as bisphosphonates or parathyroid hormone. Key messages d
Lifetime FFxs were reported by 41% (90/221) of the patients with ISM.
d
Male sex, high levels of the bone resorption marker sCTX, low hip BMD, absence of UP, and alcohol intake were independent predictors of future FFxs in patients presenting with ISM.
d
The MastFx score, a prediction model consisting of these 5 characteristics, showed good accuracy to determine the risk of new FFxs in patients with ISM.
d
QFracture, a risk scoring tool validated for FFxs in the general population, was not useful in the prediction of fracture risk in patients with ISM.
REFERENCES 1. Valent P, Akin C, Escribano L, Fodinger M, Hartmann K, Brockow K, et al. Standards and standardization in mastocytosis: consensus statements on diagnostics, treatment recommendations and response criteria. Eur J Clin Invest 2007;37: 435-53. 2. Valent P, Akin C, Arock M, Brockow K, Butterfield JH, Carter MC, et al. Definitions, criteria and global classification of mast cell disorders with special reference to mast cell activation syndromes: a consensus proposal. Int Arch Allergy Immunol 2012;157:215-25.
J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 6
3. Barete S, Assous N, de Gennes C, Grandpeix C, Feger F, Palmerini F, et al. Systemic mastocytosis and bone involvement in a cohort of 75 patients. Ann Rheum Dis 2010;69:1838-41. 4. Van der Veer E, van der Goot W, de Monchy JG, Kluin-Nelemans HC, van Doormaal JJ. High prevalence of fractures and osteoporosis in patients with indolent systemic mastocytosis. Allergy 2012;67:431-8. 5. Escribano L, Alvarez-Twose I, Sanchez-Mu~noz L, Garcia-Montero A, Nu~nez R, Almeida J, et al. Prognosis in adult indolent systemic mastocytosis: a long-term study of the Spanish Network on Mastocytosis in a series of 145 patients. J Allergy Clin Immunol 2009;124:514-21. 6. Alvarez-Twose I, Gonzalez de Olano D, Sanchez-Mu~noz L, Matito A, EstebanLopez MI, Vega A, et al. Clinical, biological, and molecular characteristics of clonal mast cell disorders presenting with systemic mast cell activation symptoms. J Allergy Clin Immunol 2010;125:1269-78. 7. Rossini M, Zanotti R, Bonadonna P, Artuso A, Caruso B, Schena D, et al. Bone mineral density, bone turnover markers and fractures in patients with indolent systemic mastocytosis. Bone 2011;49:880-5. 8. Guillaume N, Desoutter J, Chandesris O, Merlusca L, Henry I, Georgin-Lavialle S, et al. Bone complications of mastocytosis: a link between clinical and biological characteristics. Am J Med 2013;126(75):e1-7. 9. Seitz S, Barvencik F, Koehne T, Priemel M, Pogoda P, Semler J, et al. Increased osteoblast and osteoclast indices in individuals with systemic mastocytosis. Osteoporos Int 2013;24:2325-34. 10. Chiappetta N, Gruber B. The role of mast cells in osteoporosis. Semin Arthritis Rheum 2006;36:32-6. 11. Rajgopal R, Bear M, Butcher MK, Shaughnessy SG. The effects of heparin and low molecular weight heparins on bone. Thromb Res 2008;122:293-8. 12. Theoharides TC, Alysandratos KD, Angelidou A, Delivanis DA, Sismanopoulos N, Zhang B, et al. Mast cells and inflammation. Biochim Biophys Acta 2012;1822: 21-33. 13. Theoharides TC, Boucher W, Spear K. Serum interleukin-6 reflects disease severity and osteoporosis in mastocytosis patients. Int Arch Allergy Immunol 2002;128: 344-50. 14. Brockow K, Akin C, Huber M, Metcalfe DD. IL-6 levels predict disease variant and extent of organ involvement in patients with mastocytosis. Clin Immunol 2005;115:216-23. 15. Sticherling M, Brasch J, Br€uning H, Christophers E. Urticarial and anaphylactoid reactions following ethanol intake. Br J Dermatol 1995;132:464-7. 16. Longley J, Duffy TP, Kohn S. The mast cell and mast cell disease. J Am Acad Dermatol 1995;32:545-61. 17. Murphy M, Watson R. Antihistamine syrups and alcohol. Pediatr Dermatol 2001; 18:539. 18. National Clinical Guideline Centre (UK). Osteoporosis: fragility fracture risk: osteoporosis: assessing the risk of fragility fracture. London: Royal College of Physicians (UK); 2012. 19. Kanis JA, Johnell O, Oden A, Sembo I, Redlund-Johnell I, Dawson A, et al. Long-term risk of osteoporotic fracture in Malm€o. Osteoporos Int 2000;11: 669-74.
VAN DER VEER ET AL 1421
20. Van Staa TP, Dennison EM, Leufkens HGM, Cooper C. Epidemiology of fractures in England and Wales. Bone 2001;29:517-22. 21. Ismail AA, Pye SR, Cockerill WC, Lunt M, Silman AJ, Reeve J, et al. Incidence of limb fracture across Europe: results from the European Prospective Osteoporosis Study (EPOS). Osteoporos Int 2002;13:565-71. 22. Hedstr€ om EM, Svensson O, Bergstr€om U, Michno P. Epidemiology of fractures in children and adolescents. Acta Orthop 2010;81:148-53. 23. Cooper C, Dennison EM, Leufkens HG, Bishop N, van Staa TP. Epidemiology of childhood fractures in Britain: a study using the general practice research database. J Bone Miner Res 2004;19:1976-81. 24. Horny HP, Akin C, Metcalfe DD, Escribano L, Bennett JM, Valent P, et al. Mastocytosis (Mast cell disease). In: Swerdlow P, Campo P, Harris P, Jaffe P, Pileri P, Stein P, editors. World Health Organization (WHO) classification of tumours. Pathology & genetics. Tumours of haematopoietic and lymphoid tissues, Vol 2. Lyon (France): IARC Press; 2008. pp. 54-63. 25. Kanis JA, Oden A, Johnell O, Jonsson B, De LC, Dawson A. The burden of osteoporotic fractures: a method for setting intervention thresholds. Osteoporos Int 2001;12:417-27. 26. Arends S, Spoorenberg A, Bruyn GA, Houtman PM, Leijsma MK, Kallenberg CG, et al. The relation between bone mineral density, bone turnover markers, and vitamin D status in ankylosing spondylitis patients with active disease: a crosssectional analysis. Osteoporos Int 2011;22:1431-9. 27. Granerus G, Lonnqvist B, Nystrand J, Roupe G. Serum tryptase measured with B12 and G5 antibody-based immunoassays in mastocytosis patients and its relation to histamine turnover. Br J Dermatol 1998;139:858-61. 28. Keyzer JJ, Wolthers BG, Muskiet FA, Kauffman HF, Groen A. Determination of N tau-methylhistamine in plasma and urine by isotope dilution mass fragmentography. Clin Chim Acta 1981;113:165-73. 29. Keyzer JJ, Wolthers BG, Breukelman H, Kauffman HF, de Monchy JG. Determination of N tau-methylimidazoleacetic acid (a histamine metabolite) in urine by gas chromatography using nitrogen-phosphorus detection. Clin Chim Acta 1982; 121:379-87. 30. Oosting E, Keyzer JJ, Wolthers BG, Scholtis RJ. Age dependent normal values of histamine and histamine metabolites in human urine. Agents Actions 1988;23: 307-10. 31. Rubin KH, Friis-Holmberg T, Hermann AP, Abrahamsen B, Brixen K. Risk assessment tools to identify women with increased risk of osteoporotic fracture: complexity or simplicity? A systematic review. J Bone Miner Res 2013;28: 1701-17. 32. Kanis JA. Diagnosis of osteoporosis and assessment of fracture risk, review. Lancet 2002;359:1929-36. 33. Ivaska KK, Gerdhem P, Vaananen HK, Akesson K, Obrant KJ. Bone turnover markers and prediction of fracture: a prospective follow-up study of 1040 elderly women for a mean of 9 years. J Bone Miner Res 2010;25:393-403. 34. Vasikaran S, Eastell R, Bruyere O, Foldes AJ, Garnero P, Griesmacher A, et al. IOF-IFCC Bone Marker Standards Working Group. Markers of bone turnover for the prediction of fracture risk and monitoring of osteoporosis treatment: a need for international reference standards. Osteoporos Int 2011;2:391-420.
From aLaboratory Medicine, bRheumatology and Clinical Immunology, cAllergy, and d the GRIAC Research Institute, University of Groningen, University Medical Center Groningen. Disclosure of potential conflict of interest: The authors declare that they have no relevant conflicts of interest. Received for publication September 13, 2013; revised April 3, 2014; accepted for publication May 8, 2014. Available online June 27, 2014. Corresponding author: Eveline van der Veer, PhD, Department of Laboratory Medicine, HPC-EA40, University Medical Center Groningen, PO Box 30.001, 9700 RB Groningen, The Netherlands. E-mail: [email protected]. 0091-6749/$36.00 Ó 2014 American Academy of Allergy, Asthma & Immunology http://dx.doi.org/10.1016/j.jaci.2014.05.003
Abbreviations used AUC: Area under the curve BMD: Bone mineral density BMI: Body mass index BTM: Bone turnover marker CVinter: Interassay coefficient of variation FFx: Fragility fracture FRAX: Fracture Risk Assessment Tool HR: Hazard ratio ISM: Indolent systemic mastocytosis MH: Methylhistamine MIMA: Methylimidazole acetic acid ROC: Receiver operating characteristic sCTX: Serum type I collagen C-telopeptide UP: Urticaria pigmentosa
Indolent systemic mastocytosis (ISM) is defined by the presence of abnormal mast cells outside the skin, principally in the bone marrow.1 Mast cells produce and release a large number of different mediators, such as histamine, prostaglandin D2, platelet-activating factor, proinflammatory cytokines, leukotrienes C4 and D4, chemokines, and tryptase.2 Symptoms and signs are caused by release of these mast cell mediators, local accumulation of mast cells, or both. Manifestations vary strongly in nature and severity and can include skin and gastrointestinal symptoms, anaphylaxis, musculoskeletal pain, and neuropsychiatric symptoms. Fragility fractures (FFxs), osteoporosis, or both occur in approximately half of the patients with ISM.3-9 Notably, the prevalence of FFxs was high in men, even in those less than 50 years of age.4 Mast cell mediators are branded to have deleterious effects on bone metabolism.10,11 In addition, activated mast cells release the proinflammatory cytokines IL-1, IL-6, and TNF-a,12 which are important regulators of bone resorption.10 Some patients with ISM exhibit increased serum IL-6 levels, and a correlation with disease severity was found in a group of patients with different types of mastocytosis.13,14 Furthermore, alcohol triggers mast cell degranulation.15-17 Its intake can be associated with a number of adverse reactions in patients with ISM. Apart from toxicological effects, intolerance occurs and might manifest clinically as flushing and, in rare cases, as generalized urticaria and anaphylactic reactions after ingestion. In the general population multiple tools have been developed to assess a patient’s fracture risk based on models that integrate age and sex with additional risk factors (eg, low body mass index [BMI], previous FFxs, underlying diseases, use of corticosteroids, smoking, and excessive alcohol intake), with or without the use of bone mineral density (BMD) at the femoral neck. Recently, the National Institute of Clinical Excellence’s guidelines 1413
1414 VAN DER VEER ET AL
recommended assessing bone health by using one of the 2 validated scoring systems, namely the Fracture Risk Assessment Tool (FRAX) and QFracture.18 However, these fracture risk scoring systems do not take into account specific risk factors associated with FFxs in the younger population. FRAX is limited to persons older than 40 years. Even though QFracture includes persons between 30 and 40 years of age, the score is less reliable because the number of fractures in the data set is small for this age group. In addition, the prevalence of multiple fractures in patients with ISM3,4 is much higher than the prevalence found in European population studies.19-23 Therefore the aim of the present study was to investigate the predictive value of certain patient characteristics at the time of diagnosis for the occurrence of future FFxs in patients with ISM. By using these characteristics, an easyto-use prediction model, named the MastFx score, was developed and tested to identify patients with ISM at risk of new FFxs.
METHODS Study population Between 1981 and August 2012, ISM was diagnosed in 228 consecutive patients at the referral centre for mastocytosis of the University Medical Center Groningen. The diagnosis was established by fulfilling the criteria of the World Health Organization Classification of Mastocytosis. Patients with aggressive forms of systemic mastocytosis were excluded. Smoldering systemic mastocytosis, a provisional subentity of ISM with detectable B findings (criteria used for classification of systemic mastocytosis) caused by the mast cell disease process, was excluded as well.24 Fracture data were collected from medical records and lateral radiographs of the thoracolumbar spine combined with a questionnaire sent to all patients. FFxs, also known as low-level (or ‘‘low-energy’’) trauma fractures, were defined as fractures that result from mechanical forces that would not ordinarily result in fractures.25 In cases in which it was not possible to discriminate between high- or low-energy trauma circumstances, a telephone call was made to clarify. Fracture data of 7 (3%) of the 228 patients could not be retrieved, and these patients were excluded. BMD of the lumbar spine (anterior-posterior projection at L1-L4) and hip (total proximal femur) were measured with dual energy x-ray absorptiometry (DXA) (Hologic QDR Discovery, Waltham, Mass). BMD T-scores were calculated by using the National Health and Nutrition Examination Survey reference database. Osteoporosis was defined according to World Health Organization guidelines as a BMD T-score of 22.5 or less, and osteopenia was defined as a BMD T-score of between 21.0 and 22.5. In addition, data regarding osteosclerosis, the presence of urticaria pigmentosa (UP), and anaphylaxis were collected (definitions, details, or both were published previously4). For the predictor analysis, 40 of the 221 patients were excluded because of treatment for osteoporosis (eg, bisphosphonates or parathyroid hormone) before ISM diagnosis (n 5 28), sex change (n 5 1), fracture or surgery within the last 6 months before diagnosis (n 5 2), or missing BMD and bone turnover marker (BTM) data (n 5 9, Fig 1). The Medical Ethical Review Board of the University Medical Center Groningen declared that the study was performed in accordance with regulations of the review board for publication of subjects’ data.
Biochemical markers BTM was studied by means of assessment of the formation markers bonespecific alkaline phosphatase, procollagen type 1 N-terminal peptide, and osteocalcin and the resorption marker serum type I collagen C-telopeptide (sCTX).26 Bone-specific alkaline phosphatase levels were measured by using ELISA (Metra Biosystems, Mountain View, Calif; interassay coefficient of variation [CVinter], 5.5%), procollagen type 1 N-terminal peptide levels were measured by using an RIA (Orion Diagnostica, Espoo, Finland; CVinter, 9.0%), osteocalcin levels were measured by using an immunoradiometric assay (BioSource Europe S.A, Nivelles, Belgium; CVinter, 9.4%), and sCTX
J ALLERGY CLIN IMMUNOL DECEMBER 2014
levels were measured by using an electrochemiluminescence immunoassay (Elecsys 2010; Roche, Mannheim, Germany; CVinter, 10.8%). Serum samples were stored within 1 hour after venipuncture at 2208C until analysis. BTM Z-scores, the number of SDs from the normal mean corrected for age and sex, were calculated by using a Dutch reference group (200 men and 350 women) checked for serum 25-hydroxyvitamin D levels of greater than 50 nmol/L, as well as for the absence of osteoporosis (BMD T-score, >22.5) after 50 years of age. Z-scores were calculated as follows: (BTM value of individual patient 2 Mean BTM value of matched 10-year cohort of healthy reference group)/SD of this matched reference cohort. Tryptase levels were determined by using the B12 assay (Pharmacia UniCAP Tryptase; Thermo Fisher Scientific/Pharmacia and Upjohn, Uppsala, Sweden; CVinter, 5.8%).27 Reference values for healthy subjects are from Pharmacia and Upjohn, showing a geometric mean level of 3.8 mg/L and an upper 95th percentile of 11.4 mg/L. Urine samples were collected after an overnight fast, discarding the first voiding after wakening, to measure methylhistamine (MH) and methylimidazole acetic acid (MIMA) values. For 24 hours before urine collection, patients were asked to refrain from histamine-rich foods and drinks, such as sauerkraut, canned fish, yogurt, and wine. The MH value was determined by using an isotope-dilution mass fragmentographic method.28 MIMA was determined as described previously,29 with some modifications using isotope dilution mass fragmentography. For MH and MIMA excretion, mean 6 SD values of an apparently healthy population were 101 6 33 mmol/mol (range, 50-154 mmol/mol; CVinter, 6.8%) and 1.3 6 0.3 mmol/mol (range, 0.9-1.9 mmol/mol; CVinter, 4.2%) creatinine, respectively.30
Statistical analysis Statistical analysis was performed with IBM SPSS Statistics 20 software (SPSS, Chicago, Ill). Results were expressed as means 6 SDs or medians (ranges) for parametric and nonparametric data, respectively. Predictor analysis for the occurrence of FFxs was performed by using univariate and multivariate Cox regression with backward Wald inclusion of variables with a P value of .20 or less in univariate analysis. The P value for stepwise removal was .10, and P values of less than .05 were considered statistically significant. For continuous variables with significant outcomes in univariate analysis, relevant cutoff points were determined according to clinical applicability and checked by using receiver operating characteristic (ROC) curves. These categorized variables were used in multivariate analysis. The final prediction model, named the MastFx score, was obtained by rounding the regression coefficients of the multivariate model. ROC analysis was used to determine the performance of this prediction model (area under the curve [AUC] < 0.70 was interpreted as poor, 0.70 < AUC < 0.80 was interpreted as moderate, 0.80 < AUC < 0.90 was interpreted as good, and AUC > 0.90 was interpreted as high accuracy). Finally, the MastFx scores were compared with risk scores calculated with one of the National Institute of Clinical Excellence–recommended validated scoring tools, the QFracture-2012 algorithm (version 1.0; ClinRisk Ltd, Leeds, United Kingdom). For patients with ISM aged less than 30 years (n 5 22), the QFracture risk was calculated with an age of 30 years.
RESULTS FFxs Information about lifetime fractures was retrieved for 221 (97%) patients. In total, 389 lifetime fractures occurred in 127 patients. Of these fractures, 125 were high-energy trauma fractures, and 264 were FFxs. Before the ISM diagnosis was made, 54 patients reported 139 FFxs. After the ISM diagnosis, 56 patients reported 125 new FFxs during a median follow-up of 5.4 years (range, 0.4-15.3 years; Fig 2, A), with a 5- and 10-year fracture-free survival of 77% 6 3% and 69% 6 4%, respectively. Forty patients were excluded from the predictor analysis to avoid bias of factors influencing fracture risk, BMD, and/or BTM (Fig 1). In the remaining 181 patients with ISM, 27 reported 40
VAN DER VEER ET AL 1415
J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 6
FIG 1. Flowchart of handling of fracture data from patients with ISM.
FFxs before ISM diagnosis, and 43 reported 88 FFxs after diagnosis (Fig 2, B). Median follow-up (5.3 years; range, 0.4-15.3 years) and 5- and 10-year fracture-free survival (78% 6 4% and 71% 6 4%) were comparable with those of the total group.
Patient characteristics Characteristics at the time of diagnosis of the 181 patients with ISM included in the predictor analysis are presented in Table I. Mean age was 46 6 13 years, and 59% of the patients were female. Only 3 women had a BMI of less than 19 kg/m2. Another 14 women weighed less than 60 kg, although they had normal BMI. Current smoking or smoking stopped within the last 10 years was reported by 64% of the patients, with differences in smoking habits between male and female subjects (58% vs 38%, P 5.04). Seventy-six percent reported social alcohol intake, but only 8% used 3 or more units per day. Thirty-eight patients with ISM reported mild-to-moderate adverse reactions after past intake of alcohol. Median duration between first symptoms and diagnosis was 7.5 years (range, 0.1-58 years). UP was present in 76% of the patients, 43% experienced anaphylactic reactions, and 35% used at least 1 antimediator drug.
Osteoporosis of the lumbar spine and hip was detected in 14% and 0.6%, respectively, and 15% of the patients had a previous FFx. The 43 patients with new FFxs during follow-up had normal BMD (n 5 12 [28%]), osteopenia (n 5 20 [47%]), or osteoporosis (n 5 11 [26%]) at the time of diagnosis.
Predictors Patients with new FFxs after ISM diagnosis were older; were more often male; had more frequent anaphylactic reactions; had UP less often; had higher MIMA, osteocalcin, and sCTX levels; had lower hip BMD scores; and reported more frequent alcohol intake at the moment of diagnosis (Table II). For clinical applicability, relevant cutoff points were determined (Table II) and checked by using ROC curves (data not shown). Multivariate Cox regression analysis showed that male sex (hazard ratio [HR], 2.04), a sCTX Z-score of 11.0 or greater (HR, 2.63), a hip BMD T-score of 21.0 or less (HR, 2.19), absence of UP (HR, 2.05), and alcohol intake (HR, 3.45) at the time of ISM diagnosis were independent predictors of future FFxs (Table III). The fracture-free survival curves for each of the predictors are shown in Fig 3. Further investigation revealed that only 33% of
1416 VAN DER VEER ET AL
J ALLERGY CLIN IMMUNOL DECEMBER 2014
FIG 2. Lifetime FFxs in patients before and/or after the diagnosis of ISM. A, All patients with ISM (n 5 221). B, Patients with ISM used for the predictor analysis (n 5 181).
the patients without UP were less than 50 years of age compared with 66% in the UP group (mean age, 53.8 6 11.6 vs 43.9 6 12.1 years; P < .001). Forty-seven percent of the patients without UP were male compared with 40% in the UP group (P 5 .39).
MastFx score The fracture prediction model, including male sex, high sCTX levels, low hip BMD, absence of UP, and alcohol intake at the time of diagnosis, was used to calculate the MastFx score for each patient by using the regression coefficients of the multivariate model. The accuracy of this model to predict which patients will have fragile fractures was good, with an AUC of 0.80 (95% CI, 0.73-0.88). Comparable results were found if, for clinical applicability, the regression coefficients were rounded to 1 point (AUC, 0.80; 95% CI, 0.72-0.87). Fracture-free survival curves for all patients receiving a MastFx score of 0 to 5 are shown in Fig 4. Patients with a score of 0 (n 5 22) or 1 (n 5 53) had a 5- and 10-year fracture-free survival of 93%; a score of 2 points (n 5 58) resulted in 5- and 10year fracture-free survivals of 78% and 73%, respectively; and a score of 3 points (n 5 35) resulted in 5- and 10-year fracturefree survivals of 65% and 46%, respectively. A small group with a MastFx score of 4 (n 5 12) or 5 (n 5 1) had a 5-year fracture-free survival of 14%. The QFracture tool was used to compare the MastFx score with an already existing and validated scoring system for the risk of FFxs in the general population. The median 5- and 10 year risks of 181 patients with ISM for osteoporotic fracture (hip, spine, wrist, or shoulder) were 0.5% (0.1% to 5.4%) and 1.2% (0.2% to 13.6%)
and the risks for hip fracture were only 0.0% (0.0 to 2.9%) and 0.1% (0.0 to 8.7%), respectively. The risk scores had poor accuracy, with AUC values of 0.60 (95% CI, 0.50-0.71) for major osteoporotic fractures and 0.66 (95% CI, 0.56-0.75) for hip fractures. However, QFracture is not recommended for patients who have experienced an osteoporotic fracture. Excluding 27 patients with ISM with a history of FFxs, the median 5- and 10-year risks of osteoporotic fracture (hip, spine, wrist, or shoulder) were 0.5% (0.1% to 5.4%) and 1.2% (0.3% to 13.6%), respectively, and the median 5- and 10-year risks of hip fracture were only 0.0% (0.0% to 1.9%) and 0.1% (0.0% to 5.5%), respectively. The accuracy for the QFracture score was poor, and only the prediction of hip fracture reached significance (data not shown).
DISCUSSION Patients with ISM have a higher risk of fractures and osteoporosis compared with the general population. In the present study only a quarter of the patients who experienced new fractures after diagnosis had osteoporosis at the time of diagnosis. Thus BMD alone does not have sufficient power to predict the risk of new fractures and to make the clinical decision of whom to treat preventively in this patient group. A prediction model was developed using patient characteristics at the time of diagnosis to be able to make a better estimate of the risk of fractures. The characteristics included in this model, named the MastFx score, are male sex, high sCTX levels, low hip BMD, absence of UP, and alcohol intake, and these 5 parameters can easily be obtained in daily clinical practice. The MastFx score shows good accuracy
VAN DER VEER ET AL 1417
J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 6
TABLE I. Characteristics at the time of diagnosis of 181 patients with ISM Patient characteristics Age (y) Male sex Length (cm) Weight (kg) BMI (kg/m2) Smoking Alcohol use Mastocytosis characteristics Disease duration (y) UP Anaphylactic shock Tryptase (mg/L) MH (mmol/mol creatinine) MIMA (mmol/mol creatinine) Bone characteristics Patients with FFxs No. of FFxs Osteosclerosis Osteocalcin (mg/L) Osteocalcin Z-score (SD) BALP (U/L) BALP Z-score (SD) PINP (mg/L) PINP Z-score (SD) sCTx (pg/mL) sCTX Z-score (SD) LS BMD (g/cm2) LS BMD T-score (SD) Hip BMD (g/cm2) Hip BMD T-score (SD) Comorbidity Patients with comorbidity Diabetes mellitus type 2 Hypothyroidism Hyperthyroidism Hypogonadism Hyperparathyroidism Hypophosphatemia Pulmonary disease Epilepsy Malignancy Rheumatic diseases Crohn disease Other diseases Alcohol abuse Drug abuse Current use of antimediator drug Patients using antimediators H1- and/or H2-antihistamines Cromoglicic acid Oral corticosteroids Inhaled corticosteroids
46 6 13 74 (41%) 174 6 10 79 6 15 26.3 6 4.4 82 (46%) 135 (75%) 7.5 (0.1 to 58) 138 (76%) 77 (43%) 27 (4.1 to 296) 257 (70 to 2554) 3.2 (0.5 to 21.6) 27 (15%) 40 10 (6%) 12.3 (3.6 to 37.2) 20.12 (22.66 to 4.15) 21.5 (4.6 to 62.1) 1.22 (22.09 to 8.55) 39.9 (15.9 to 135) 20.04 (22.10 to 6.62) 195 (10 to 797) 20.42 (22.03 to 6.35) 0.96 6 0.15 20.97 6 1.39 0.93 6 0.13 20.35 6 0.99
_50 y Age >
77 (42%)
Weight 20 mg/L MH >167 mmol/mol creatinine MIMA >1.9 mmol/mol creatinine
(8%) (1.6%) (19%) (30%) (16%) (21%)
122 (67%) 141 (78%) 148 (82%)
Osteocalcin Z-score >12.0 SD
10 (6%)
BALP Z-score >12.0 SD
57 (32%)
PINP Z-score >12.0 SD
12 (7%)
sCTX Z-score >12.0 SD
14 (8%)
Osteoporosis LS
25 (14%)
Osteoporosis hip
1 (0.6%)
48 (27%) 5 4 3 2 1 1 19 1 3 4 2 12 3 2 63 (35%) 55 (31%) 11 (6%) 4 (2.2%) 7 (3.9%)
Values are presented as means 6 SDs, medians (ranges), or numbers (percentages) unless otherwise indicated. BALP, Bone-specific alkaline phosphatase; LS, lumbar spine; PINP, procollagen type 1 N-terminal peptide.
(AUC, 0.80) to determine the risk of future FFxs in patients with ISM. The predictive accuracy of the MastFx score in our data set is much better than that of QFracture, which is a validated tool for assessing the fracture risk in persons aged between 30 and 99
years without a history of osteoporotic fracture.18 For comparison with the MastFx model, patients with ISM and a history of FFxs were included, but the QFracture accuracy was not altered substantially without them. The QFracture model is based on databases of the United Kingdom population. Because no validated
1418 VAN DER VEER ET AL
J ALLERGY CLIN IMMUNOL DECEMBER 2014
TABLE II. Univariable Cox regression to identify parameters that are related to new FFxs occurring after the ISM diagnosis 95% CI Variable
95% CI
HR
Lower
Upper
P value
Cutoff
HR
Lower
Upper
P value
1.015
1.069
.002
_50 y
1 1.995
1.087
3.660
.026
< _4 mmol/mol creatinine >4 mmol/mol creatinine
1 0.998
0.540
1.844
.995
_1.0 SD
1.823
0.918
3.622
.086
_1.0 SD
1 2.929
1.562
5.494
.001
>21.0 SD < _21.0 SD
1 2.496
1.366
4.563
.003
Age
Y
1.042
Sex
Female Male cm kg kg/m2
1 2.554 1.014 1.011 1.016
Absent Present Alcohol use Absent Present Disease duration y Anaphylactic reaction Absent Present UP Absent Present Tryptase mg/L MH (mmol/mol creatinine)/10 MIMA mmol/mol creatinine
1 1.191 1 4.688 0.982 1 2.033 2.602 1 1.003 1.007 1.084
FFxs
Osteocalcin Z-score
n/Patient Absent Present SD
BALP Z-score PINP Z-score
1.379 0.983 0.99 0.954
4.732 1.046 1.031 1.082
.003 .379 .306 .619
0.649
2.184
.573
1.448 0.946
15.18 1.018
.010 .322
1.094 1.423
3.780 4.759
.025 .002
0.998 0.998 1.002
1.009 1.015 1.174
.263 .110 .044
1.366 1 1.648 1.272
0.991
1.882
.057
0.810 1.010
3.352 1.601
.168 .041
SD SD
0.981 1.166
0.806 0.923
1.194 1.473
.851 .199
sCTX Z-score
SD
1.217
1.026
1.442
.024
LS BMD T-score Hip BMD T-score
SD SD
0.883 0.669
0.715 0.484
1.091 0.924
.250 .015
Osteoporosis*
Absent Present Absent Present Absent Present
1 1.747 1 1.369 1 1.119
Length Weight BMI Smoking
Comorbidity Antimediator use
0.880
3.469
.111
0.723
2.594
.335
0.643
1.946
.691
Values in boldface indicate statistical significance. BALP, Bone-specific alkaline phosphatase; PINP, procollagen type 1 N-terminal peptide; LS, lumbar spine. *Defined as a lumbar spine, hip, or both BMD T score of 22.5 SD or less.
TABLE III. Multivariate Cox regression to identify independent predictors of new FFxs after ISM diagnosis 95% CI
Male sex _11.0 sCTX Z-score > _21.0 Hip BMD T-score < UP absence Alcohol use
HR
Lower
Upper
P value
B
MastFx score
2.043 2.632 2.187 2.047 3.445
1.045 1.278 1.128 1.074 1.016
3.996 5.424 4.240 3.899 11.688
.037 .009 .021 .029 .047
0.715 0.968 0.782 0.716 1.237
1 1 1 1 1
Values in boldface indicate statistical significance. B, Regression coefficient.
risk calculation tools were available for the general Dutch population, we used the country for which the epidemiology of fractures in the general population most closely approximates our country. Because 30% of the patients with ISM were aged less than 40 years, no comparison was made with the FRAX scores. A recent review compared 6 different fracture risk assessment
tools, including QFracture and FRAX, which were externally validated in a population-based setting with proper methodological quality. The authors found that no tool performed consistently better than others (ie, AUC values mostly ranged between 0.6 and 0.8), indicating that all the tools are modest predictors of fractures.31
J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 6
VAN DER VEER ET AL 1419
FIG 3. FFx-free survival curves of the 5 independent predictors: male sex, sCTX Z-score of 11.0 or greater, hip BMD T-score of 21.0 or less, absence of UP, and use of alcohol.
In the general population older age is a well-known risk factor for the development of fractures. In our mastocytosis cohort 24% (54/221) of patients had FFxs at the time of diagnosis, which increased to 41% at the study’s closing date (median follow-up, 5.4 years). Although this indicates the relevance of age, this parameter was not identified as an independent predictor in our MastFx model. Compared with the general population, fractures occurred more frequently in young men with ISM and postmenopausal women with ISM and not in premenopausal women with ISM, as we discussed in our previous study.4 An interesting finding was that the presence of UP protected against fractures.
Further exploration of this finding is needed. In our cohort 49 (27%) of the 181 patients were women less than 50 years of age, 45 of whom had UP. Although only 2 of these patients had new FFxs after ISM diagnosis, no significant interactions were found between age or sex and UP. As expected, high sCTX levels and low hip BMD were associated with FFxs, which is in agreement with earlier publications.32-34 Surprisingly, the mast cell mediators were statistically not selected in the prediction model. Previously, we have shown that higher MH levels were associated with a higher risk of osteoporotic manifestations (osteoporotic fractures and
1420 VAN DER VEER ET AL
FIG 4. FFx-free survival curves per MastFx score (0-5) using a prediction model consisting of 5 patient characteristics.
osteoporosis combined).4 In the present study only MIMA levels had a significant positive association with new osteoporotic fractures in univariate analysis. Perhaps urinary histamine metabolite levels reflect mast cell burden in bone marrow to only a minor extent. Another possibility might be that the deleterious effect of local exposure from the mast cell mediators on the bone is of greater importance than the effect of circulating mediators, as indicated by levels measured in serum and urine. A third explanation might be that the histamine metabolites are lesser risk factors when BTM is taken into account. In different European cohorts the percentage of patients with ISM and FFxs varied from 5% to 37%.3-5,7 These differences can probably be explained by differences in age of the included patients and in the methods used to explore the history and circumstances of lifetime fractures. The 24% history of FFxs in our cohort with a mean age of 48 years is in line with the 21% rate of vertebral fractures in patients with ISM of the same age reported by Italian authors.7 The fracture rate was substantially lower in a younger Spanish cohort.5 In addition, there might be differences in antimediator treatment strategies between the European cohorts. In this study we found no significant difference in future fracture rate between patients with and without antimediator drug use at the time of ISM diagnosis. We were not able to collect data on treatment duration and doses used. Therefore further exploration of the clinical relevance of antimediator drugs on the occurrence of fractures is needed.
J ALLERGY CLIN IMMUNOL DECEMBER 2014
A limitation of this study is the exclusion of 30% of the fractures occurring during the follow-up period by the exclusion of 40 patients, including the 28 who had already been treated for osteoporosis. This method of patient selection might have caused a loss of association with new fractures. Arguments for excluding patients receiving medication for osteoporosis were that these patients were known to be at risk and that measures to reduce the risk of fractures were already taken before the ISM diagnosis was made. In summary, we presented a discriminative and easy-to-use prediction model for the occurrence of FFxs using 5 patient characteristics. Lifetime FFxs were reported by 41% (90/221) of the patients with ISM. Because FFxs can have a devastating effect on the physical and mental condition of the individual patient, appropriate risk assessment is needed. Calculation of the FFx risk should be an important component in the management of patients presenting with ISM. The risk of an individual patient is indicated by the MastFx score, which is calculated as the sum of the following 5 characteristics (1 point each): male sex, high levels of _11.0), low hip BMD (T-score, < _21.0), absence sCTX (Z-score, > of UP, and alcohol intake. On the basis of our prediction model with scores ranging from 0 to 5, patients with ISM with a MastFx score of 2 or greater have a high risk of FFxs. Additional analyses showed that the MastFx score is more useful than an existing fracture risk calculation tool developed for the older general population. Efforts should be made by caretakers to optimize bone quality with adequate vitamin D and calcium intake and lifestyle changes in all patients with ISM. Alcohol cessation in these patients is highly recommended because drinking is a modifiable risk factor for FFxs in patients with ISM. By using the MastFx score, it is possible to distinguish patients with ISM at high fracture risk from those at low fracture risk. These high-risk patients will probably benefit from an early start of therapeutic intervention with drugs, such as bisphosphonates or parathyroid hormone. Key messages d
Lifetime FFxs were reported by 41% (90/221) of the patients with ISM.
d
Male sex, high levels of the bone resorption marker sCTX, low hip BMD, absence of UP, and alcohol intake were independent predictors of future FFxs in patients presenting with ISM.
d
The MastFx score, a prediction model consisting of these 5 characteristics, showed good accuracy to determine the risk of new FFxs in patients with ISM.
d
QFracture, a risk scoring tool validated for FFxs in the general population, was not useful in the prediction of fracture risk in patients with ISM.
REFERENCES 1. Valent P, Akin C, Escribano L, Fodinger M, Hartmann K, Brockow K, et al. Standards and standardization in mastocytosis: consensus statements on diagnostics, treatment recommendations and response criteria. Eur J Clin Invest 2007;37: 435-53. 2. Valent P, Akin C, Arock M, Brockow K, Butterfield JH, Carter MC, et al. Definitions, criteria and global classification of mast cell disorders with special reference to mast cell activation syndromes: a consensus proposal. Int Arch Allergy Immunol 2012;157:215-25.
J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 6
3. Barete S, Assous N, de Gennes C, Grandpeix C, Feger F, Palmerini F, et al. Systemic mastocytosis and bone involvement in a cohort of 75 patients. Ann Rheum Dis 2010;69:1838-41. 4. Van der Veer E, van der Goot W, de Monchy JG, Kluin-Nelemans HC, van Doormaal JJ. High prevalence of fractures and osteoporosis in patients with indolent systemic mastocytosis. Allergy 2012;67:431-8. 5. Escribano L, Alvarez-Twose I, Sanchez-Mu~noz L, Garcia-Montero A, Nu~nez R, Almeida J, et al. Prognosis in adult indolent systemic mastocytosis: a long-term study of the Spanish Network on Mastocytosis in a series of 145 patients. J Allergy Clin Immunol 2009;124:514-21. 6. Alvarez-Twose I, Gonzalez de Olano D, Sanchez-Mu~noz L, Matito A, EstebanLopez MI, Vega A, et al. Clinical, biological, and molecular characteristics of clonal mast cell disorders presenting with systemic mast cell activation symptoms. J Allergy Clin Immunol 2010;125:1269-78. 7. Rossini M, Zanotti R, Bonadonna P, Artuso A, Caruso B, Schena D, et al. Bone mineral density, bone turnover markers and fractures in patients with indolent systemic mastocytosis. Bone 2011;49:880-5. 8. Guillaume N, Desoutter J, Chandesris O, Merlusca L, Henry I, Georgin-Lavialle S, et al. Bone complications of mastocytosis: a link between clinical and biological characteristics. Am J Med 2013;126(75):e1-7. 9. Seitz S, Barvencik F, Koehne T, Priemel M, Pogoda P, Semler J, et al. Increased osteoblast and osteoclast indices in individuals with systemic mastocytosis. Osteoporos Int 2013;24:2325-34. 10. Chiappetta N, Gruber B. The role of mast cells in osteoporosis. Semin Arthritis Rheum 2006;36:32-6. 11. Rajgopal R, Bear M, Butcher MK, Shaughnessy SG. The effects of heparin and low molecular weight heparins on bone. Thromb Res 2008;122:293-8. 12. Theoharides TC, Alysandratos KD, Angelidou A, Delivanis DA, Sismanopoulos N, Zhang B, et al. Mast cells and inflammation. Biochim Biophys Acta 2012;1822: 21-33. 13. Theoharides TC, Boucher W, Spear K. Serum interleukin-6 reflects disease severity and osteoporosis in mastocytosis patients. Int Arch Allergy Immunol 2002;128: 344-50. 14. Brockow K, Akin C, Huber M, Metcalfe DD. IL-6 levels predict disease variant and extent of organ involvement in patients with mastocytosis. Clin Immunol 2005;115:216-23. 15. Sticherling M, Brasch J, Br€uning H, Christophers E. Urticarial and anaphylactoid reactions following ethanol intake. Br J Dermatol 1995;132:464-7. 16. Longley J, Duffy TP, Kohn S. The mast cell and mast cell disease. J Am Acad Dermatol 1995;32:545-61. 17. Murphy M, Watson R. Antihistamine syrups and alcohol. Pediatr Dermatol 2001; 18:539. 18. National Clinical Guideline Centre (UK). Osteoporosis: fragility fracture risk: osteoporosis: assessing the risk of fragility fracture. London: Royal College of Physicians (UK); 2012. 19. Kanis JA, Johnell O, Oden A, Sembo I, Redlund-Johnell I, Dawson A, et al. Long-term risk of osteoporotic fracture in Malm€o. Osteoporos Int 2000;11: 669-74.
VAN DER VEER ET AL 1421
20. Van Staa TP, Dennison EM, Leufkens HGM, Cooper C. Epidemiology of fractures in England and Wales. Bone 2001;29:517-22. 21. Ismail AA, Pye SR, Cockerill WC, Lunt M, Silman AJ, Reeve J, et al. Incidence of limb fracture across Europe: results from the European Prospective Osteoporosis Study (EPOS). Osteoporos Int 2002;13:565-71. 22. Hedstr€ om EM, Svensson O, Bergstr€om U, Michno P. Epidemiology of fractures in children and adolescents. Acta Orthop 2010;81:148-53. 23. Cooper C, Dennison EM, Leufkens HG, Bishop N, van Staa TP. Epidemiology of childhood fractures in Britain: a study using the general practice research database. J Bone Miner Res 2004;19:1976-81. 24. Horny HP, Akin C, Metcalfe DD, Escribano L, Bennett JM, Valent P, et al. Mastocytosis (Mast cell disease). In: Swerdlow P, Campo P, Harris P, Jaffe P, Pileri P, Stein P, editors. World Health Organization (WHO) classification of tumours. Pathology & genetics. Tumours of haematopoietic and lymphoid tissues, Vol 2. Lyon (France): IARC Press; 2008. pp. 54-63. 25. Kanis JA, Oden A, Johnell O, Jonsson B, De LC, Dawson A. The burden of osteoporotic fractures: a method for setting intervention thresholds. Osteoporos Int 2001;12:417-27. 26. Arends S, Spoorenberg A, Bruyn GA, Houtman PM, Leijsma MK, Kallenberg CG, et al. The relation between bone mineral density, bone turnover markers, and vitamin D status in ankylosing spondylitis patients with active disease: a crosssectional analysis. Osteoporos Int 2011;22:1431-9. 27. Granerus G, Lonnqvist B, Nystrand J, Roupe G. Serum tryptase measured with B12 and G5 antibody-based immunoassays in mastocytosis patients and its relation to histamine turnover. Br J Dermatol 1998;139:858-61. 28. Keyzer JJ, Wolthers BG, Muskiet FA, Kauffman HF, Groen A. Determination of N tau-methylhistamine in plasma and urine by isotope dilution mass fragmentography. Clin Chim Acta 1981;113:165-73. 29. Keyzer JJ, Wolthers BG, Breukelman H, Kauffman HF, de Monchy JG. Determination of N tau-methylimidazoleacetic acid (a histamine metabolite) in urine by gas chromatography using nitrogen-phosphorus detection. Clin Chim Acta 1982; 121:379-87. 30. Oosting E, Keyzer JJ, Wolthers BG, Scholtis RJ. Age dependent normal values of histamine and histamine metabolites in human urine. Agents Actions 1988;23: 307-10. 31. Rubin KH, Friis-Holmberg T, Hermann AP, Abrahamsen B, Brixen K. Risk assessment tools to identify women with increased risk of osteoporotic fracture: complexity or simplicity? A systematic review. J Bone Miner Res 2013;28: 1701-17. 32. Kanis JA. Diagnosis of osteoporosis and assessment of fracture risk, review. Lancet 2002;359:1929-36. 33. Ivaska KK, Gerdhem P, Vaananen HK, Akesson K, Obrant KJ. Bone turnover markers and prediction of fracture: a prospective follow-up study of 1040 elderly women for a mean of 9 years. J Bone Miner Res 2010;25:393-403. 34. Vasikaran S, Eastell R, Bruyere O, Foldes AJ, Garnero P, Griesmacher A, et al. IOF-IFCC Bone Marker Standards Working Group. Markers of bone turnover for the prediction of fracture risk and monitoring of osteoporosis treatment: a need for international reference standards. Osteoporos Int 2011;2:391-420.
