JBMR

ORIGINAL ARTICLE

Quantitative and Qualitative Changes of Bone in Psoriasis and Psoriatic Arthritis Patients Roland Kocijan,1,2 Matthias Englbrecht,1 Judith Haschka,1,2 David Simon,1 Arnd Kleyer,1 Stephanie Finzel,1 €rgen Rech,1* Sebastian Kraus,1 Heinrich Resch,2 Christian Muschitz,2 Klaus Engelke,3 Michael Sticherling,4 Ju 1 and Georg Schett * 1

Department of Internal Medicine 3 and Institute of Clinical Immunology, University of Erlangen-Nuremberg, Erlangen, Germany St. Vincent Hospital, Medical Department II, The VINFORCE Study Group, Academic Teaching Hospital of Medical University of Vienna, Vienna, Austria 3 Institute of Medical Physics, Erlangen-Nuremberg, Erlangen, Germany 4 Department of Dermatology, University of Erlangen-Nuremberg, Erlangen, Germany 2

ABSTRACT Psoriatic arthritis (PsA) is a chronic inflammatory disease characterized by periarticular bone loss and new bone formation. Current data regarding systemic bone loss and bone mineral density (BMD) in PsA are conflicting. The aim of this study was to evaluate bone microstructure and volumetric BMD (vBMD) in patients with PsA and psoriasis. We performed HR-pQCT scans at the ultradistal and periarticular radius in 50 PsA patients, 30 psoriasis patients, and 70 healthy, age- and sex-related controls assessing trabecular bone volume (BV/TV), trabecular number (Tb.N), inhomogeneity of the trabecular network, cortical thickness (Ct.Th), and cortical porosity (Ct.Po), as well as vBMD. Trabecular BMD (Tb.BMD, p ¼ 0.021, 12.0%), BV/TV (p ¼ 0.020, –11.9%), and Tb.N (p ¼ 0.035, 7.1%) were significantly decreased at the ultradistal radius and the periarticular radius in PsA patients compared to controls. In contrast, bone architecture of the ultradistal radius and periarticular radius was similar in patients with psoriasis and healthy controls. Duration of skin disease was associated with low BV/TV and Tb.N in patients with PsA. These data suggest that trabecular BMD and bone microstructure are decreased in PsA patients. The observation that duration of skin disease determines bone loss in PsA supports the concept of subclinical musculoskeletal disease in psoriasis patients. © 2015 American Society for Bone and Mineral Research. KEY WORDS: PSORIATIC ARTHRITIS; PSORIASIS; BONE MINERAL DENSITY; BONE MICROSTRUCTURE; COMPUTED TOMOGRAPHY

Introduction

P

soriasis is an immunologically mediated, chronic inflammatory disorder of the skin. It is characterized by increased cytokine production in the skin due to a combination of genetic and environmental factors.(1,2) Recent evidence suggests that about 30% of patients with psoriasis have psoriatic arthritis (PsA) in hospital clinics and about 11% in the community.(3–5) Psoriasis and PsA are characterized by tissue infiltrates of activated T-cells in the skin and the synovial tissue and local increase in production of tumor necrosis factor (TNF)-alpha, interleukin (IL)17, and IL-23.(6–9) T cells and cytokine-activated resident cells have been defined as effective triggers of osteoclast differentiation and bone loss in inflammatory conditions by inducing the expression of receptor activator of nuclear factor kappa B ligand (RANKL).(10) This link between the immune system and the bone may be relevant for psoriasis and PsA, because the gestalt of cytokine expression in these diseases promotes a pro-osteoclastogenic

milieu and may promote bone loss.(11) PsA is characterized by synovitis and enthesitis, which can change the local physiologic bone homeostasis. Local bone loss reflected by bone erosions occurs similarly in patients with rheumatoid arthritis and PsA. In addition, PsA is also characterized by sites of new bone formation, leading to a typical shape of bone erosions as well as the formation of bony spurs.(12) However, data on systemic bone loss, reflected by decreased bone mineral density (BMD) in PsA and psoriasis, are conflicting(13–18) and have so far been confined to evaluate total BMD by dual-energy X-ray absorptiometry (DXA), whereas detailed analyses of the bone structure in PsA and psoriasis patients have not been undertaken to date. Highresolution peripheral quantitative computed tomography (HRpQCT) allows a separate analysis of trabecular BMD (Tb.BMD) and cortical BMD (Ct.BMD) as well as bone microstructure and geometry of the peripheral bone.(19,20) Because fracture risk does not solely depend on quantitative total BMD but also on bone quality related to the microstructure of cortical and

Received in original form August 26, 2014; revised form March 11, 2015; accepted March 26, 2015; accepted manuscript online March 31, 2015. Address correspondence to: Georg Schett, MD, Department of Internal Medicine 3, Rheumatology and Immunology; University of Erlangen-Nuremberg; Ulmenweg 18, Erlangen, D-91054, Germany. E-mail: [email protected] *JR and GS contributed equally to this work. Journal of Bone and Mineral Research, Vol. 30, No. 10, October 2015, pp 1775–1783 DOI: 10.1002/jbmr.2521 © 2015 American Society for Bone and Mineral Research

1775

trabecular bone, more detailed knowledge on the bone architecture in PsA and psoriasis is important, given the fact that an increased vertebral and nonvertebral fracture incidence in psoriasis and PsA patients compared to healthy controls has been reported.(18,21,22) Thus, the primary objective of this study was to search for deterioration in volumetric BMD (vBMD) and in the microstructure of trabecular and cortical bone at non-periarticular and periarticular sites in patients with PsA. The secondary objective was to find evidence for a potential skin-bone axis in psoriasis by testing whether patients with skin disease develop bone changes independently of arthritis.

Patients and Methods

28 (DAS28) was recorded in PsA and the Psoriasis Area and Severity Index (PASI) was evaluated in psoriasis patients. Current and previous treatment with conventional disease-modifying anti-rheumatic drugs (DMARDs; methotrexate, leflunomide) and biological DMARDs (tumor necrosis factor-alpha inhibitors) was recorded. In addition, the use of high-dose oral glucocorticoids (>5 mg equivalent to prednisolone daily for 3 months) and low-dose oral glucocorticoids (5 mg equivalent to prednisolone daily for 3 months) was documented. Low-trauma fractures, defined as fractures acquired without any identifiable trauma or following a minor injury that would not typically lead to a bone fracture in an individual, were also recorded.(24) Clinical low-trauma fractures were assessed by questionnaire. No radiographs of the vertebra were performed. All fractures were self-reported, peripheral fractures.

Patients and study design In this prospective observational study, 50 patients with nonaxial PsA (26 women, 24 men), 30 patients with skin psoriasis (13 women, 17 men), and 70 healthy controls (35 women, 35 men) were enrolled. PsA patients had to fulfill the Classification Criteria for Psoriatic Arthritis (CASPAR) criteria to be enrolled into this study.(23) Hence, patients must have inflammatory articular disease and at least three of the following five points: (1) current psoriasis or personal or family history for psoriasis; (2) psoriatic nail dystrophy; (3) absence of rheumatoid factor; (4) current or previous dactylitis; and (5) periarticular new bone formation. Patients with axial spondyloarthritis, rheumatoid arthritis, or connective tissue diseases were excluded from analysis. Patients were included in the psoriasis group if psoriatic skin disease was present, which had to be diagnosed by a dermatologist. Exclusion criteria for the psoriasis group were a history or current features of joint disease. All patients received rheumatologic assessment to exclude involvement of the peripheral joints, the spine, and the entheses. Enthesial involvement was additionally excluded by high-resolution ultrasound examination. Healthy, age-related control subjects were randomly selected from an HR-pQCT database of a prospectively collected population of healthy men and women. These individuals were recruited and measured at the Department of Internal Medicine 3 via an announcement for prevention of osteoporosis. In all these individuals concomitant diseases were ruled out by taking patients’ history and by clinical examination. Subjects receiving antiosteoporotic treatment including antiresorptive therapy (bisphosphonates, denosumab, raloxifene, strontium ranelate) or osteoanabolic therapy (teriparatide) were generally excluded from this study. Patients were recruited at the Department of Internal Medicine 3 and the Department of Dermatology of the University of Erlangen-Nuremberg. The local ethics committee and the National Radiation Safety Agency (Bundesamt fuer Strahlenschutz) approved this study. The study was performed in accordance with the Declaration of Helsinki. Subjects were only enrolled in the study after signing an informed consent form.

Image acquisition We analyzed vBMD, bone microstructure of the trabecular and cortical bone, and bone geometry by HR-pQCT (Xtreme-CT €ttisellen, Switzerland) using the manufacscanner; Scanco, Bru turer’s standard in vivo protocol. For cortical porosity (Ct.Po, %) the advanced cortical analysis protocol with a fully automated cortical compartment segmentation was used.(25) Measurements were performed at the ultradistal and periarticular radius, close to the joint gap. Measurements at the ultradistal radius were carried out 9.5 mm proximal to the reference line, as described.(25) For the periarticular radius the reference line was set through the edge of the medial part of the radio-carpal joint.(26) The volumes of interest for both sites are shown in Fig. 1. Measurements were performed at the clinically more affected arm in PsA and usually the right arm in patients with psoriasis and healthy subjects. For measurement, the forearm was immobilized in a carbon fiber cast. A total of 110 slices (voxel size 80 mm, nominal resolution 140 mm, 60 kVp effective energy, 900 mA) were carried out for both measuring sites. All measurements were performed with the latest standard analysis

Clinical data We assessed demographic data such as age, sex, weight, height, body mass index (BMI), and smoking status (current or previous smoker). Disease-specific variables including duration of skin and joint disease, high-sensitive C-reactive protein level, as well as autoantibodies (rheumatoid factors [RF] and anti-cyclic citrullinated protein antibodies [ACPA]) were determined in patients with PsA and psoriasis. Moreover, Disease Activity Score

1776

KOCIJAN ET AL.

Fig. 1. Scout view of the volumes of interest at the radius. Ultradistal radius (green dotted line) and periarticular radius (red continuous line).

Journal of Bone and Mineral Research

software (version 6.0) as described.(26) Daily measurements with the standard phantom were performed to determine the longitudinal stability of the Xtreme-CT measurements, according the reference manual (XtremeCT, User’s Guide, Revision 5.05). Reproducibility of measurements was not systematically assessed in this study. However, short-term, intermediate-term, and long-term single-center reproducibility for the standard in vivo protocol of the Xtreme-CT scanner is high, with a precision error of less than 5%.(27)

Image analysis vBMD (mg hydroxyapatite [HA]/cm3) including total BMD (average BMD, mg HA/cm3), trabecular BMD (Tb.BMD, mg HA/cm3), and cortical BMD (Ct.BMD, mg HA/cm3) were determined. Trabecular bone microstructure including trabecular bone volume fraction (BV/TV, %), trabecular number (Tb.N, 1/mm), trabecular thickness (Tb.Th, mm), trabecular separation (Tb.Sp, mm), and inhomogeneity of the network (Tb.1/N.SD, mm) were assessed in all subjects. Moreover, cortical thickness (Ct.Th, mm) and cortical porosity (Ct.Po, %) were analyzed. Cortical perimeter (Ct.Pm, mm) and bone area (mm2), reflecting bone geometry were additionally evaluated. Because of thin cortices at the periarticular radius, cortical parameters were not analyzed at this site.

Statistical analysis Differences in vBMD, microstructure, and bone geometry between PsA patients, psoriasis patients, and healthy controls, respectively, were determined by analysis of covariance (ANCOVA) accounting for the influence of sex. Sidak post hoc tests for pairwise comparisons accounting for multiple tests were performed if ANCOVA results suggested significant differences between subgroups. In patients with PsA or psoriasis, we evaluated the relation of bone parameters to independent variables (ie, sex, current therapy with glucocorticoids, biologics, or DMARDs, smoking status, age, BMI, and duration of psoriasis) in distal (BMD, BV/TV, Tb.N, CtTh) and ultradistal (BMD, BV/TV, Tb.N) regions by partial least squares regression (PLSR) of main effects. We choose this method because of the sample size and its limitations with respect to regular linear regression. PLSR is an approach that combines variance explanation and the characterization of factors contributing to variance explanation by independent variables. The number of latent factors was limited with respect to the difference of variance explained when another factor was added; ie, adjusted R2 had to be increased by 1% at least. This procedure implies that the first factor in the solution explains more variance of the dependent variable than each of the subsequent factors. Results of factor solution were only interpreted, if at least 25% of the dependent variable’s variance was explained and one or more factor loadings were exceeding 0.50. For this analysis, we used Python Essentials extension package for IBM SPSS, version 21. Because of the numerical imbalance in body weight, smoking, and sex among PsA patients, psoriasis patients, and controls a linear regression model including average BMD, BV/TV, Tb.N, and Ct.Th was set up. All independent variables were included into the regression model by entering all variables at a single step. A dummy coding procedure was used in order to include multicategorical variables properly. To test associations between low-traumatic fractures and exposure to glucocorticoids, DMARDs, and biologicals, a chi square test (x2) was performed in patients with PsA.

Journal of Bone and Mineral Research

Results Demographic data Patients with PsA, psoriasis, and healthy controls were balanced with respect to age (50.9  12.8, 50.7  13.2, and 51.5  13.5 years, respectively), gender (52%, 43%, and 50% women, respectively), height (172.4  8.6, 172.6  9.8, and 171.0  9.9 cm, respectively), and BMI (27.9  5.1, 27.4  4.8, and 26.2  4.8, respectively) without any significant differences. However, weight was significantly lower for controls compared to PsA (83.5  16.9 versus 76.6  14.9 kg, p ¼ 0.020) but comparable between controls and patients with psoriasis (81.9  16.3 versus 76.6  14.9 kg, p ¼ 0.120). Disease duration of PsA was 8.0  7.3 years. Duration of skin disease was significantly longer in PsA than in psoriasis patients (21.7  14.1 versus 14.3  15.0 years, p ¼ 0.028). Rheumatoid factor was found in only one patient with PsA and no patient with psoriasis or controls. All PsA patients, psoriasis patients, and controls were ACPA-negative. C-reactive protein was low and comparable between PsA and psoriasis (4.4  7.1 versus 4.0  5.3 mg/L, p ¼ 0.766). Disease activity in PsA was generally low to moderate (DAS28: 3.0  1.5). Low-trauma fractures occurred in 14% (n ¼ 7, 6 women, 1 man) of PsA patients and in none of the psoriasis patients or controls. All fractures were selfreported, peripheral fractures. Glucocorticoid use (>5 mg equivalent to prednisolone daily for 3 months) and low-dose oral glucocorticoid use (5 mg equivalent to prednisolone daily for 3 months) was uncommon in PsA (4% and 16%, respectively). No patient in the psoriasis or in the control group received glucocorticoids. Conventional and biological DMARD intake was frequent in patients with PsA (60% and 60%, respectively), whereas only a few patients with psoriasis received conventional and biological DMARDs (6% and 3%, respectively). Details are shown in Table 1. There were no significant associations between low-traumatic fractures and glucocorticoids (x ¼ 3.488, p ¼ 0.175), current DMARDs (x ¼ 2.243, p ¼ 0.134), number of DMARDs (x ¼ 1.872, p ¼ 0.599), current biologicals (x ¼ 0.028, p ¼ 0.868), and number of biologicals (x ¼ 4.281, p ¼ 0.639), respectively.

vBMD and microstructure in PsA In the first part of the analysis we compared patients with PsA to healthy controls. ANCOVA results suggested significant differences regarding Tb.BMD, BV/TV, Tb.N, and inhomogeneity of the trabecular network. Corresponding Sidak post hoc comparisons showed significant differences between PsA and controls regarding Tb.BMD (162.1  39.8 versus 181.5  39.0 mg HA/ cm3, p ¼ 0.021, –12.0%), trabecular bone volume (BV/TV, 0.135  0.03 versus 0.151  0.03, p ¼ 0.020, –11.9%), and trabecular number (Tb.N, 1.98  0.31 versus 2.12  .27, p ¼ 0.035, –7.1%), as well as an increased trabecular separation (450  100 versus 409  70, p ¼ 0.028, 12.0%) in PsA patients compared to controls (see Figs. 2 and 3). Ct.BMD (827.7  76.2 versus 855.1  53.2 mg HA/cm3, –3.2%) and total vBMD (317.6  64.3 versus 336.3  49.8 mg HA/cm3, –5.6%) were nonsignificantly decreased in PsA compared to controls (see Table 2). No significant differences were found regarding cortical parameters including cortical thickness, cortical porosity, and cortical perimeter. At the periarticular radius, close to the joint, PsA patients had a lower Tb.BMD (171.7  34.8 versus 185.6  34.9, p ¼ 0.056, –8.1%) and BV/TV (0.143  0.03 versus 0.155  0.03, p ¼ 0.056, –8.4%) compared to controls. All data are summarized in Table 2.

BONE MICROSTRUCTURE IN PSORIATIC ARTHRITIS

1777

Table 1. Patient Characteristics

Demographic characteristics Patients (n) Sex (men/women) Age (years) Height (cm) Weight (kg) BMI (kg/m2) Current or previous smoker (%) Disease-specific characteristics Duration of PsA (years) Duration of skin psoriasis (years) ACPA positive, n (%) RF positive, n (%) C-reactive protein (mg/L) DAS28 score (units) PASI Low trauma fractures, n (%) Treatment modalities Current DMARDs, n (%) Current and previous DMARDs, n Current biologics, n (%) Current and previous biologics, n Current glucocorticoids, n (%)b Current LD glucocorticoids, n (%)c Current antiresorptive drugs, n (%)

PsA

Psoriasis

Controls

50 24/26 50.9  12.8 172.4  8.6 83.5  16.9 27.9  5.1 59

30 17/13 50.7  13.2 172.6  9.8 81.9  16.3 27.4  4.8 62

70 35/35 51.5  13.5 171.0  9.9 76.6  14.9‡ 26.2  4.8 29

8.0  7.3 21.7  14.1 0 (0) 1 (2) 4.4  7.1 3.0  1.5 7.5  8.5 7 (14)

n.a. 14.3  15.0a 0 (0) 0 (0) 4.0  5.3 n.a. 7.8  9.5 0 (0)

n.a. n.a. 0 (0) 0 (0) n.a. n.a. n.a. 0 (0)

30 (60) 1.2  0.8 30 (60) 1.4  1.6 2 (4) 8 (16) 0 (0)

2 (6) 0.3  0.5a 1 (3) 0.1  0.4a 0 (0) 0 (0) 0 (0)

n.a. n.a. n.a. n.a. 0 (0) 0 (0) 0 (0)

Demographic and disease specific characteristics of patients with PsA, psoriasis, and healthy controls. Values are mean  SD, absolute values, or percentages. PsA ¼ psoriatic arthritis; BMI ¼ body mass index; ACPA, anti-citrullinated protein antibodies; RF ¼ rheumatoid factor; DAS28 ¼ Disease Activity Score 28; PASI ¼ Psoriasis Area and Severity Index; DMARD, disease modifying anti-rheumatic drug; n.a. ¼ not applicable; LD ¼ low-dose. a Significant difference to PsA, p < 0.05. b 5 mg prednisolone for at least 3 months. c LD glucocorticoids: 5 mg prednisolone for at least 3 months.

Fig. 2. HR-pQCT scans of representative patients with psoriatic arthritis (A), psoriasis (B), and healthy control (C). Ultradistal radius, axial view.

1778

KOCIJAN ET AL.

Journal of Bone and Mineral Research

Fig. 3. Bone microstructure in psoriatic arthritis, psoriasis, and healthy controls. Reconstruction of high-resolution peripheral quantitative computed tomography scans of male patients with psoriatic arthritis (A, red) and psoriasis (B, green), and healthy controls (C, purple). Upper row: 3D reconstruction, 20 slices, axial view. Lower row: 3D reconstruction, 30 slices, axial/sagittal view.

vBMD and microstructure in psoriasis In a second step we compared patients with psoriasis to healthy controls. At the ultradistal and periarticular radius we did not find any significant differences regarding vBMD, microstructure, or geometry between the two subgroups, suggesting a very similar bone phenotype in patients with psoriasis without arthritis and healthy subjects.

Regression models for bone microstructural parameters Only the regression models for BMD, BV/TV, and Tb.N at the periarticular radius were found to fulfill the predefined criteria of variance explanation, suggesting an important contribution of independent variables. For BMD we obtained a three-factor solution with age loading highest on factor 1 (–0.68) and absence of therapy with biologics on factor 2 (–0.69). For BV/TV (R2adj ¼ 0.43), we obtained two latent factors importantly contributing to variance explanation, with age (–0.53) and duration of skin psoriasis (–0.55) loading highest on factor 1 and absence of current therapy with biologics (–0.64) on factor 2. For Tb.N, PLSR again suggested a two latent factor solution with gender (0.62) and duration of skin psoriasis (–0.60) loaded highest on factor 1, whereas glucocorticoid therapy (–0.54) loaded highest on factor two. Data are shown in Table 3.

Regression model for demographic data We next performed a regression model to find out which parameters are independently determining BV/TV and Tb.N of the ultradistal radius. The “diagnosis of PsA” (BV/TV: b ¼ –0.205, t ¼ –2.221, p ¼ 0.028; Tb.N: b –0.228, t ¼ –2.393, p ¼ 0.018) as well as female gender (BV/TV: b ¼ –0.418, t ¼ –4.589, p < 0.001;

Journal of Bone and Mineral Research

Tb.N: b ¼ –0.254, t ¼ –2.708, p ¼ 0.008) were independently associated with low BV/TV and low Tb.N. Similar results were found in the periarticular radius, showing an independent association between trabecular microstructure and the diagnosis of PsA as well as female gender. In contrast, no independent associations were found between the microstructure parameters and smoking, body weight, or diagnosis of psoriasis (Table 4).

Discussion In the present study we show that PsA is associated with significantly decreased trabecular vBMD and deterioration of trabecular bone microstructure. No significant differences were observed regarding cortical BMD and microstructure including cortical thickness and cortical porosity in PsA patients, which contrast with findings in rheumatoid arthritis (RA).(26) Patients with psoriatic skin disease without arthritis showed no significant trabecular bone loss compared to healthy controls. Duration of skin disease was, next to higher age, a key factor associated with trabecular bone loss in patients with PsA. Furthermore, as expected, female sex was associated with trabecular bone loss in PsA, patients with psoriasis, and healthy controls. We recently reported decreased trabecular and cortical microstructure in patients with RA.(26) Whereas bone loss and decreased BMD are well described in RA, data on BMD in PsA are conflicting.(13–18) In contrast to RA, PsA is not only characterized by bone loss, but also by sites of new bone formation, which typically affect the insertion sites of the tendons and ligaments at the periarticular bone. Unfortunately, previous studies

BONE MICROSTRUCTURE IN PSORIATIC ARTHRITIS

1779

Table 2. Bone Microstructure in PsA and PsO ANOVA

Ultradistal radius Bone geometry Bone area (mm2) Ct.Pm (mm) vBMD Average BMD (HA/cm3) Tb.BMD (HA/cm3) Ct.BMD (HA/cm3) Bone microstructure BV/TV (%) Tb.N (mm–1) Tb.Th (mm) Tb.Sp (mm) Inhomogeneity (mm) Ct.Th (mm) Ct.Po (%) Periarticular radius Bone geometry Bone area (mm2) Ct.Pm (mm) vBMD Tb.BMD (HA/cm3) Ct.BMD (HA/cm3) Bone microstructure BV/TV (%) Tb.N (mm–1) Tb.Th (mm) Tb. Sp (mm) Inhomogeneity (mm)

PsA versus Ctrl

PsO versus Ctrl

PsA

PsO

Ctrl

F

p

p

Diff (%)

p

Diff (%)a

327  93 78.6  12.7

324  92 77.3  2.4

324  82 77.0  10.9

0.49 1.08

0.609 0.340

 

1.0 2.0

 

–0.1 0.4

317  64 162  39 827  76

335  64 175  43 847  69

336  49 181  39 855  53

1.62 3.74 2.67

0.200 0.026 0.072

 0.021 

5.5 12.0 3.3

 0.708 

0.4 3.2 0.9

13.5  3 1.98  0.31 68  10 450  100 200  60 807  200 2.4  1.2

14.6  4 2.10  0.31 69  10 419  90 178  70 844  200 2.4  1.1

15.1  3 2.12  0.27 71  10 409  70 176  60 846  160 2.4  1.2

3.76 3.33 1.25 3.66 2.16 0.68 0.01

0.025 0.038 0.288 0.028 0.119 0.505 0.983

0.020 0.035  0.023   

11.9 7.1 1.5 9.1 12.0 4.8 0.0

0.699 0.944  0.814   

3.4 1.0 2.9 2.4 1.1 0.2 0.0

462  125 96.2  14.7

464  125 95.2  13.8

460  104 95.1  12.3

0.12 0.51

0.882 0.597

 

0.4 1.1

 

1.0 0.1

171  34 

191  31 

185  34 

3.65 

0.029 

0.056 

8.1 

0.961 

3.0 

14.3  3 2.15  0.27 66  10 406  60 179  50

16.0  3.0 2.26  0.22 71  10 375  40 155  30

15.5  3.0 2.23  0.25 69  10 385  50 166  40

3.66 2.03 1.88 3.01 2.63

0.028 0.135 0.156 0.052 0.076

0.056    

8.4 3.7 4.5 5.2 7.3

0.958    

3.1 1.3 2.8 2.7 7.1

a

Bone geometry, vBMD, and microstructure by HR-pQCT in PsA, PsO, and Ctrl. ANCOVA results controlling for sex effects. Sidac post hoc tests are only shown for significant ANCOVA results. Results are mean  SD. PsA ¼ psoriatic arthritis; Ctrl ¼ healthy controls; PsO ¼ psoriasis; Diff ¼ difference; Ct ¼ cortical; Pm ¼ perimeter; BMD ¼ bone mineral density; HA ¼ hydroxyapatite; vBMD ¼ volumetric BMD; Tb ¼ trabecular; BV/TV ¼ bone volume fraction; Tb.N ¼ trabecular number; Tb.Th ¼ trabecular thickness; Tb.Sp ¼ trabecular separation; Ct.Th ¼ cortical thickness; Ct.Po ¼ cortical porosity. a Differences (%) were calculated using the following formulas: PsA versus Ctrl ¼ (PsA – Ctrl)/Ctrls; PsO versus Ctrl ¼ (PsO – Ctrl)/Ctrl.

assessing bone in psoriasis and PsA used 2D DXA, which is not able to differentiate between cortical and trabecular bone. Although total BMD was only slightly altered in patients with PsA, Tb.BMD, trabecular volume, and trabecular numbers were significantly lower in patients with PsA than in healthy controls. In contrast, no differences were found regarding cortical microstructure, which differs from the findings in RA patients.(26) These observations support data from Anandarajah and colleagues,(28) who reported significantly lower BMD in the lumbar spine, but not in the femoral neck, in DXA scans of PsA patients. Because the lumbar spine is mainly composed by trabecular bone, whereas cortical bone prevails in the femoral neck, decreased lumbar spine BMD supports the dominance of trabecular bone loss in patients with PsA. Frediani and colleagues(13) also suggested mainly qualitative alterations in the trabecular microarchitecture of patients with PsA. These data are in accordance with our results, showing primarily deterioration of trabecular, but not cortical bone microstructure in the HR-pQCT, which has the advantage to separately assess cortical and trabecular bone loss. In contrast, Zhu and colleagues(29) recently reported similar Tb.BMD and trabecular microstructure in Chinese patients with PsA and

1780

KOCIJAN ET AL.

controls. In their study, only cortical parameters such as Ct.BMD and cortical porosity differed between PsA and healthy controls. It has to be taken into account that bone microstructure in Asians is different from Caucasians, characterized by thicker cortices and fewer bony trabeculae.(30) This might be a valuable explanation for the different findings in our Caucasian patients with PsA. Next to age and sex, the duration of skin disease was associated with trabecular bone loss in PsA patients in the multivariate model. This observation suggests that the longterm exposure of the skeleton to psoriatic skin inflammation may lead to accumulation of bone damage over time. Hence, trabecular bone loss in PsA patients may not arise from local inflammatory changes in the joints but rather by exposure of the bone to the inflammatory cytokines associated with skin disease such as TNF-a and IL-17, which are both considered being highly potent in disturbing bone homeostasis. If so, psoriasis patients without joint disease may show signs of bone loss as well. Although we did not find a significant decrease of bone mass in the psoriasis patients compared to healthy controls in our study, values for Tb.BMD, trabecular number, and trabecular thickness were numerically lower in psoriasis patients than in healthy

Journal of Bone and Mineral Research

Table 3. Regression Models for Bone Microstructural Parameters BMD

Ultradistal radius Gender (male) Glucocorticoids (no) Current biologics (no) Current DMARDs (no) Smoker (no) Age BMI Duration of skin psoriasis R2 adjusted

Periarticular radius Gender (male) Glucocorticoids (no) Current biologics (no) Current DMARDs (no) Smoker (no) Age BMI Duration of skin psoriasis R2 adjusted

BV/TV

Tb.N

Ct.Th

Factor 1

Factor 2

Factor 1

Factor 2

Factor 1

Factor 1

0.13 0.08 0.49 0.27 0.46 0.59 0.39 0.37

0.16 0.41 0.57 0.36 0.54 0.01 0.20 0.49

0.58 0.17 0.28 0.31 0.08 0.56 0.10 0.46

0.35 0.26 0.55 0.27 0.58 0.04 0.30 0.42

0.80 0.01 0.12 0.24 0.08 0.79 0.19 0.33 0.19

0.50 0.07 0.16