Simultaneous quantification of bone erosions and enthesiophytes in the joints of patients with psoriasis or psoriatic arthritis - effects of age and disease duration

Healthy controls and patients with PsA were part of the Erlangen Imaging Cohort (ERIC), which prospectively assesses articular bone composition in healthy controls and patients with arthritis [17]. Three groups were analysed in a cross-sectional setting: (1) patients with PsA, (2) patients with PSO and (3) healthy controls. Patients with PsA were recruited at the Rheumatology Department of the University of Erlangen-Nuremberg. All patients with PsA had to fulfil the Classification criteria for psoriatic arthritis (CASPAR) criteria [18]. An experienced rheumatologist examined the patients for clinical signs of musculoskeletal involvement (synovitis, enthesitis, dactylitis and/or inflammatory back pain) at the time of imaging. Patients with PSO were recruited at the Dermatology Department of the University of Erlangen-Nuremberg. They were referred to the Rheumatology Department of the same institution for detailed rheumatologic analysis. As with PsA, an experienced rheumatologist examined the patients for the presence of clinical signs of musculoskeletal involvement. Patients with no current or previous evidence of arthritis, enthesitis or other manifestations of PsA were selected to undergo further evaluation. In addition, healthy controls with comparable age and sex were also investigated. Healthy controls were collected by a field campaign and had no joint pain, swelling or any other sign of inflammatory disease and no personal or family history of such disease. Healthy controls had not to have osteoporosis or any evidence of metabolic diseases such as diabetes mellitus, malabsorption or thyroid dysfunction. Also, renal and hepatic function had to be normal. Subjects receiving glucocorticoids or bisphosphonates (in the present or past) were also excluded.

The study was conducted upon approval of the local ethic committee of the University of Erlangen and with the authorisation of the National Radiation Safety Agency (Bundesamt für Strahlenschutz). Each patient provided informed consent.

Demographic data such as age, sex, body mass index (BMI) and smoking status were collected. As disease-specific parameters, disease severity of psoriasis according to the Psoriasis area severity index (PASI) [19], disease duration of PSO and PsA, involvement of nail and scalp and Health assessment questionnaire (HAQ) data were recorded. In patients with PsA the disease activity 28 score based on erythrocyte sedimentation rate (DAS28-ESR) was collected. C-reactive protein (milligrams per litre) was measured presence of rheumatoid factor and anti-cyclic citrullinated peptides antibodies (ACPA) was assessed. In addition, current medication including disease-modifying anti-rheumatic drugs (DMARDs), glucocorticoids and non-steroidal anti-inflammatory drugs (NSAIDs) was recorded.

All subjects underwent HR-pQCT scans of the dominant hand using the XtremeCT scanner (SCANCO Medical, Brütisellen, Switzerland). To reduce movement artefacts the hand was immobilised using a custom holder. The region of interest (ROI) was the metacarpal head and the phalangeal base of the metacarpophalangeal (MCP) joints 2 and 3. During the 8.4-min scanning time the MCP region was 80 slices distal to the edge of the third metacarpal head and 242 slices proximal to it were scanned. Resolution was 82 μm isotropic voxels.

Images were analysed using axial slices (N = 322). As described before [18], bone was divided into four quadrants representing the palmar (Q1), ulnar (Q2), dorsal (Q3), and radial (Q4) side, respectively. Thereby exact localization of each erosion or enthesiophyte was possible (Fig. 1). Erosions were defined as breaks in the cortical shell visible in at least two planes [20]. Enthesiophytes were identified as bony proliferations at specific anatomical sites (outside or along the insertion of the capsule). Based on previous studies we have exactly defined these anatomical sites by regions of interest (ROIs) [15‐17]. If an erosion or enthesiophyte was found in one (axial) plane, the lesion had to be confirmed in the perpendicular plane (coronal or sagittal).

For scoring the number of bone erosions and enthesiophytes per patient, the sum of the respective lesion number in the four aforementioned quadrants in the four assessed regions (metacarpal head 2, metacarpal head 3, phalangeal base 2, phalangeal base 3) was calculated. If in a single quadrant more than one lesion was found, only the largest lesion was scored for its dimension. Enthesiophyte size (millimetres) was defined as the distance between the highest surface of the bone proliferation and the original surface of the cortical bone. Volume of erosions (cubic millimetres) was measured according to a half-ellipsoid formula. Images were analysed and measured by two independent and blinded readers (DS, SK) using the open source DICOM viewer Osirix V4.1 (Rosslyn, VA, USA): 3D-image software provided by the manufacturer was used for obtaining illustrative 2D and 3D images.

In total, 203 individuals were analysed for catabolic and anabolic bone changes in their joints. Among these, 101 individuals (51 women/50 men) had a diagnosis of PsA, 55 (20 women/35 men) had PSO without clinical musculoskeletal involvement and 47 individuals were healthy controls (23 women/24 men). Mean age and sex distribution were comparable among the three groups. Body mass index was significantly higher in patients with PsA or PSO compared to healthy controls, representing the well-recognised association between the disease and obesity. Smoking habits were comparable among the three groups. Patients with PsA had a mean duration of 18.9 ± 14.8 years of skin disease and 6.4 ± 7.3 years of joint disease. Patients with PSO had comparable disease duration of 15.2 ± 15.4 years: 21% of patients with PsA and 51% of patients with PSO reported of nail disease, while the prevalence of scalp involvement was 29% in patients with PSO and 20% in patients with PsA. Skin and musculoskeletal disease activity in patients with PsA, as evaluated by the PASI (3.4 ± 5.5), DAS28-ESR (2.98 ± 1.48) and HAQ (0.8 ± 0.8) was low, most likely due to effective treatment of PsA. Further disease-specific characteristics of the patient groups are shown in Table 1.

Inter-observer agreement for the detection of bone lesions was very high. The intraclass correlation coefficient (ICC) was 0.96 for detection of erosions and 0.95 for detection of enthesiophytes, respectively. Regarding the extent of the lesions, the ICC for measuring the size of enthesiophytes was 0.94 and for measuring the volume of erosions it was 0.90.

In the PsA group 963 enthesiophytes were identified, while in the PSO group only 306 enthesiophytes were observed. The average number of enthesiophytes per patient was significantly higher in patients with PsA. In PsA, we detected 9.5 ± 6.7 enthesiophytes per patient, while 5.6 ± 3.3 enthesiophytes per patient were detected in PSO (p < 0.001) (Table 2). In both groups the majority of enthesiophytes were found at the metacarpal heads and the most affected sides were the dorsal and palmar side of the metacarpal heads, where functional entheses can be found. A significant (p < 0.001) difference between the PsA and PSO subgroup was detected in the extent of the enthesiophytes, with an increased enthesiophyte size of 6.8 ± 5.6 mm in the PsA group and 4.0 ± 2.2 mm in the PSO group.

In the PsA group 140 erosions were detected, while in the PSO group 27 erosions were detected. With respect to the number of erosions there was a significant difference between the PsA and PSO subgroup (p = 0.002) with patients with PsA having more erosions (1.4 ± 2.2) than patients with PSO (0.5 ± 0.9). The sides most affected by erosions were the radial sides of the metacarpal heads. There was a significant difference in the volume of erosions between patients with PsA and patients with PSO (p = 0.003), with higher erosion volumes in PsA (6.3 ± 9.7 mm³) than in PSO (3.3 ± 3.1 mm³).

We then compared catabolic and anabolic bone damage in patients with PsA with the ones identified in healthy controls. The number (p = 0.002) and size (p = 0.004) of bone erosions was significantly greater in patients with PsA compared to healthy controls. Furthermore, analysis of enthesiophytes showed a similar picture: again the number and size of enthesiophytes (both p < 0.001) was significantly greater in patients with PsA than in healthy controls. As already described [16], patients with PSO did not differ in the number of erosions compared to healthy controls, but in PSO the number of enthesiophytes was significantly greater (p < 0.001).

The analysis of number and extent of bone erosions in three age groups (20–40 years, 41–60 years and more than 60 years) revealed a continuous increase in erosive bone changes with age in healthy controls and in patients with PSO or PsA (Table 2, Fig. 2), suggesting that age is a key determinant of erosive damage in the joints. Nonetheless, patients with PsA were most affected by erosive bone damage among all age groups. Hence, the amount of erosive bone changes in a 20 to 40-year old patient with PsA was as high as in healthy individuals aged over 60 years. Healthy individuals and patients with PSO did not substantially differ in erosive changes. While erosions were absent in young healthy individuals and patients with PSO, they significantly increased with age over 60 years.

Age did not influence the burden of enthesiophytes in healthy controls and patients with PSO. In patients with PsA, the number of enthesiophytes moderately increased with age (20–40 years, 7.50 ± 5.65; 41–60 years, 9.09 ± 6.20, p = 0.015; > 60 years, 11.96 ± 7.71, p = 0.001). Similarly, the size of enthesiophytes increased with age in patients with PsA (20–40 years, 5.26 ± 5.56 mm; 41–60 years, 6.33 ± 4.96 mm, p = 0.006; 20–40 years, 5.26 ± 5.56 mm; > 60 years, 8.82 ± 6.70 mm, p = 0.001).

We next analysed to which extent the duration of PSO and PsA influence the burden of erosive damage in PsA. Bone erosions significantly increased with the duration of psoriatic skin disease (Fig. 3, Table 2). Hence, their numbers (2.65 ± 2.82) and volumes (6.72 ± 9.24 mm³) were significantly greater in patients with PsA with psoriasis of more than 20 years duration compared those with 11–20 years (p = 0.015 and p = 0.003, respectively) and less than 10 years disease duration (both p < 0.001). Not surprisingly, the duration of PsA also affected erosive damage (Fig. 3, Table 2): Patients with long disease duration (> 20 years) had significantly more erosions (3.50 ± 2.67), compared to those with disease duration between 11 and 20 years (p = 0.019) or less than 10 years (p = 0.015). In patients with PSO, overall erosive changes were very mild and not related to duration of psoriasis.

With respect to anabolic bone changes, we found that duration of skin disease influenced enthesiophyte burden both in patients with PSO and patients with PsA. While patients with PSO had mild increases in enthesiophytes with duration of psoriasis, this effect was more pronounced in PsA (Table 2). Furthermore, duration of PsA influenced the burden of enthesiophytes as well. Number (18.17 ± 10.93) and size (13.59 ± 9.71) of enthesiophytes were more than twofold greater in patients with PsA with disease duration of more than 20 years compared to those with shorter disease duration (11–20 years, p = 0.048; < 10 years, p = 0.009) (Table 2).

Duration of PSO significantly correlated with the number of enthesiophytes (p = 0.005), the size of enthesiophytes (p = 0.011), the number of erosions (p = 0.002) and the volume of erosions (p = 0.010). Duration of PsA significantly correlated with the number and size of enthesiophytes (both p < 0.001) and the number and volume of erosions (both p < 0.001). Most importantly the number and size of enthesiophytes but not number and volume of erosions correlated with physical function as measured by the HAQ score (p = 0.001, p = 0.002). Furthermore, the burden of erosions and enthesiophytes was higher (p = 0.028 and p < 0.001, respectively) in patients taking biological DMARDs compared to patients not taking biological DMARDs, reflecting the more severe disease course in patients starting biological DMARDs.

Using a regression model including sex, age, PASI, BMI and duration of skin disease to determine the factors that independently influence erosion and enthesiophytes numbers, respectively, the only variable that remained significant was the duration of skin disease (p = 0.031; p = 0.023). Using the same model and determining the extent of erosions and enthesiophytes underlined once more the importance of the duration of skin disease, as it was the only variable influencing the size of enthesiophytes and the volume of bone erosions (p = 0.005 p = 0.034).