A modified sharp score demonstrates disease progression in established psoriatic arthritis

Authors


Abstract

Objective

To use a modified Sharp score (MSS) to measure radiologic progression and to assess its relationship to other radiologic features, peripheral joint disease, and physical function in psoriatic arthritis (PsA).

Methods

Two sets of hand radiographs (median interval 5.75 years) in 139 patients with established PsA were scored using an MSS. Seventy-four patients had standardized clinical joint and Health Assessment Questionnaire (HAQ) scores and other radiologic features of PsA documented at baseline and followup (median interval 5 years).

Results

Radiologic damage was present in 58% of patients at baseline and 74% at followup. The median MSS and its components, erosion score and joint space abnormality score, were significantly greater at followup (P < 0.001). The median MSS progression was +1.08 units/year. There was strong correlation between MSS and clinical joint scores at baseline and followup (r = 0.72 and r = 0.81, respectively). There was weak correlation between MSS and HAQ at baseline (r = 0.29), but stronger correlation at followup (r = 0.48). There was a strong association between MSS and other characteristic radiologic features of PsA (bony proliferation, periostitis, bony ankylosis) at baseline and followup (P < 0.001). However, the presence of soft-tissue swelling on radiographs at baseline was the only radiologic parameter associated with an increased rate of change of MSS (corrected P < 0.006).

Conclusion

The MSS shows good construct validity with measures of peripheral joint involvement such as clinical joint scores and other radiologic features of PsA, and is able to demonstrate that radiologic damage is progressive beyond early disease.

INTRODUCTION

Psoriatic arthritis (PsA) is an inflammatory arthritis occurring in 7–42% of patients with psoriasis (1, 2). There is increasing evidence that challenges the traditionally held views that PsA is a benign disease. Prospective studies of PsA patient cohorts have documented the deteriorating functional status and progression of clinical joint scores and damage with increased disease duration (3–5).

Radiologic outcome is an important measure of disease progression. Early studies using the American College of Rheumatology (formerly the American Rheumatism Association) scoring system (6) showed progressive radiographic damage in patients with PsA. Gladman et al documented that the proportion of patients with ≥5 damaged joints increased from 19% to 40% over 5 years of followup (3). More recent studies of PsA have applied radiographic scoring systems used in rheumatoid arthritis (RA). Rahman et al, using the Steinbrocker method, demonstrated progression and radiographic severity in patients with PsA to be comparable with that in patients with RA (7). Harrison et al, using the Larsen score, found 22% of patients with PsA to have erosive disease at the end of 1 year (8).

The modified Sharp score (MSS) has been validated in patients with RA and shown to correlate well with the Larsen and Steinbrocker methods (9). The MSS contrasts with the Larsen and Steinbrocker methods in two key respects. First, for the MSS, joints are scored separately for erosion (scale of 0–5) and joint space abnormality (JSA; scale of 0–4), whereas erosion and JSA are incorporated within 1 scale for the Larsen and Steinbrocker methods. Second, more joints are scored in the hands and wrist for the MSS: 34 areas are scored for erosion and 36 areas for JSA compared with 30 areas (combined erosion and JSA) in the Steinbrocker and Larsen methods. Furthermore, the MSS may be more sensitive to change than the Larsen method (10).

There have been a number of adaptations of the MSS for use in PsA to accommodate the pattern of joints involved, which is different from RA. The most common adaptation has been to add the distal interphalangeal (DIP) joints to the PsA-MSS. Using this adaptation, Mease et al, in a trial comparing etanercept treatment versus placebo, were able to demonstrate significant differences in the total Sharp score within treatment groups as early as 6 and 12 months (11). Moreover, interobserver reliability was shown to be good. Kane et al showed significant increases in the modified Sharp erosion and JSA scores in early disease using the PsA-MSS (12). The modified Sharp/van der Heijde method for assessing radiologic damage in the hands and feet of patients with RA has also been adapted for use in PsA, using a 0–528 scale (13). The latter instrument was used to demonstrate that infliximab was more effective than placebo for inhibiting radiologic progression at 50 weeks in the Infliximab Multinational Psoriatic Arthritis Controlled Trial (14). In a modification of the Sharp score adapted for PsA (15) using a scale of 0–570, Gladman et al showed that adalimumab inhibited radiographic progression over 48 weeks (16).

The MSS in its various forms appears to be a well-validated instrument for studying radiologic progression in PsA, being sensitive to change and reliable. Both of these factors are important in a condition in which hand involvement is common (5). However, the relationship between the MSS and peripheral joint disease, physical function, and other characteristic radiologic features of PsA (construct validity) has not been well studied, nor has the PsA-MSS been used to study long-term radiologic progression beyond early disease.

In this study, we used the MSS to study its correlation with total peripheral joint involvement and physical function, and to assess the rate of progression in established disease. We also assessed whether other radiologic markers of PsA may identify patients at risk of accelerated disease progression as measured by the MSS.

MATERIALS AND METHODS

Radiographs.

Radiographs for 139 white patients with PsA were available for study. The radiographs had been taken between 1984 and 2003. All had been referred to a PsA clinic at the Royal National Hospital for Rheumatic Diseases, Bath, UK. All patients fulfilled the Moll and Wright criteria for diagnosis of PsA (17). In 74 of these patients, clinical data that were part of a previously published prospective study were available for correlation with radiology scores (5). Peripheral joint and Health Assessment Questionnaire (HAQ) scores (18) were assessed at baseline and at a median followup period of 5 years according to a pro forma identical to that completed in the prospective study (5). Seventy joints were scored for the presence of either synovial swelling or joint deformity not solely attributable to osteoarthritis. The joints included were the DIP of the hand, the interphalangeal (IP) of the thumbs, the proximal interphalangeal (PIP) of the hand, the metacarpophalangeal (MCP), the wrist, the elbow, the temperomandibular, the sternoclavicular, the acromioclavicular, the hip, the knee, the tibiotalar, the talocalcaneal, the midtarsal, the metatarsophalangeal, the IP of the first toe, and the remaining toes (each toe counting as 1). Therefore, it should be noted that the clinical joint score was a measure of both disease activity and damage.

Radiologic assessment.

Two sets of anteroposterior plain radiographs of the hands, with a median interval of 5.75 years (interquartile range [IQR] 4–7.4 years), were evaluated in 139 patients using an MSS. The first set of radiographs was obtained between 1980 and 2000, and the second set between 1993 and 2003. The reader of the radiographs (CB) was blind to the order of the radiographs and the clinical details of the patient. In the MSS, erosions (scale 0–5) and JSA (0–4) were scored separately. In an additional modification, ankylosis and joint space widening were assigned maximum JSA and erosion scores, respectively. Therefore, the following joints were evaluated in each hand for erosion: DIP, PIP, MCP, first carpometacarpal, trapezium, scaphoid, lunate, triquetrum, and distal radius and ulna. The following joints were evaluated in each hand for JSA: DIP, PIP, MCP, third to fifth carpometacarpal, trapezioscaphoid, lunatotriquetrum, lunatocapitate-scaphoid, radiocarpal, and radioulnar. Using our MSS, the maximum score for erosions was 210 and for JSA was 176. A comparison of the adapted radiologic scoring systems used in PsA is shown in Table 1.

Table 1. Adaptations of radiologic scores in PsA*
Scoring systemScale rangesNumber of areas scoredTotal erosion scoreTotal JSA scoreTotal score
ErosionJSAErosion and JSA combinedHandsFeet
  • *

    PsA = psoriatic arthritis; JSA = joint space abnormality; N/A = not applicable; PsA-MSS = modified Sharp score for PsA (current study).

  • Modified scoring method for PsA.

Steinbrocker0–43012N/AN/A168
PsA-MSS0–50–442 erosion, 44 JSAN/A210176386
Sharp/van der Heijde0–5 hands, 0–10 feet0–440 erosion, 40 JSA12 erosion, 12 JSA320208528

In 74 patients from the prospective study (5), there was additional scoring of 2 sets of radiographs for other radiologic features, some of which have been reported as being characteristic of PsA. These features were the presence or absence of soft-tissue swelling, periarticular osteoporosis (OP), bony ankylosis, periostitis, osteolysis, and bony proliferation.

Ten radiographs were evaluated by 2 different observers (JR, EK) to test interobserver agreement. Ten radiographs were evaluated by the same observer (JR) at 2 sittings, 1 month apart, to test intraobserver agreement. Intraclass correlation coefficients (ICCs) showed good agreement between the combined scores for both interobserver correlation (ICC 0.993, 95% confidence interval [95% CI] 0.981–0.997) and intraobserver correlation (ICC 0.99, 95% CI 0.991–0.999).

Statistical analysis.

The Shapiro-Wilk test was used to assess whether the data were normally distributed. Differences between paired nonparametric data sets were analyzed with Wilcoxon's signed ranks test. Differences between unpaired nonparametric data sets were analyzed using Mann-Whitney U tests. Correlation between peripheral joint scores, HAQ scores, and radiologic scores was made using Spearman's rank correlation coefficient. Bland and Altman plots were generated from the inter- and intraobserver data, and Cronbach's alpha ICCs were calculated. P values less than 0.05 were considered statistically significant. Where appropriate, Bonferroni correction was applied for multiple testing.

Ethical approval for the study was given by the Bath regional ethics committee, and informed written consent was obtained.

RESULTS

Demographic data.

Of the 139 patients with PsA studied, there were 66 men and 73 women (mean ± SD age 45 ± 13.4 years). The median disease durations at baseline and followup were 5 years (IQR 2–15 years) and 12 years (IQR 7–22 years), respectively. Thirty-eight patients had spinal disease.

Radiologic progression.

At baseline, 80 patients (58%) had evidence of radiologic damage, and this increased to 103 patients (74%) at followup. Of those with damage at baseline, 80% progressed with an increase in the combined PsA-MSS. In contrast, of those without damage at baseline, only 40% progressed.

Progression of MSS, clinical joint score, and HAQ score.

A summary of clinical and radiologic parameters at baseline and followup is shown in Table 2. The MSS and the individual components (erosion and JSA) were significantly greater at followup (for all 3 scores, P < 0.001). There was also a significant progression of clinical joint score and an increase in HAQ score with time. The median progression in MSS was +1.08 units/year (lower quartile = 0 units/year, upper quartile = 3.27 units/year). The data on MSS progression were not normally distributed. However, to create some data comparative with other studies, we also calculated the mean ± SD progression for MSS, which was +2.85 ± 5.22 units/year.

Table 2. Summary of MSS and clinical data at baseline and followup*
 BaselineFollowupP
  • *

    Values are the number (interquartile range). MSS = modified Sharp score; N/A = not applicable; JSA = joint space abnormality; HAQ = Health Assessment Questionnaire.

  • N = 139.

  • N = 74.

Median disease duration, years5.0 (2.0–15.0)12.0 (7.0–22.0)N/A
Median erosion score1.0 (0.0–10.0)5.0 (0.0–28.0)< 0.001
Median JSA score2.0 (0.0–15.0)7.0 (0.0–32.0)< 0.001
Median total MSS4.0 (0.0–31.0)14.0 (2.0–60.0)< 0.001
Clinical joint scores6.5 (2.2–17.0)12.0 (5.2–24.0)< 0.001
HAQ scores0.4 (0.0–1.0)0.5 (0.3–1.4)0.050

Correlation between erosion and JSA scores, and between PsA-MSS and peripheral joint and HAQ scores.

There was strong correlation between erosion and JSA at baseline and followup (r = 0.83 and r = 0.86, respectively).

There was strong correlation between MSS and peripheral joint scores at both baseline and followup (r = 0.72 and r = 0.81, respectively). There was weak correlation between MSS and HAQ scores at baseline (r = 0.29), and moderate correlation at followup (r = 0.48).

Other radiologic features of PsA.

There was a strong association between MSS and the presence of certain radiologic features of PsA (periarticular OP, bony proliferation, periostitis, and bony ankylosis) at baseline and followup (P < 0.001) (Table 3). The associations were similarly strong with both components of the MSS (erosion score and JSA; data not shown). There was a weaker association with radiologic evidence of soft-tissue swelling that was only significant at baseline (P < 0.05)

Table 3. Relationship between MSS and other radiologic features of PsA*
 BaselineFollowup
Radiologic featureNo.Median combined score (IQR)Mann-Whitney U testPP correctedNo.Median (IQR)Mann-Whitney U testPP corrected
  • *

    MSS = modified Sharp score; PsA = psoriatic arthritis; IQR = interquartile range.

Soft-tissue swelling  3290.0080.048  3090.307 
 No540.00 (0.00–0.70)   580.29 (0.00–1.82)   
 Yes200.56 (0.08–1.48)   130.42 (0.15–1.18)   
Osteoporosis  287< 0.001< 0.006  295< 0.001< 0.006
 No480.00 (0.00–0.22)   460.11 (0.00–0.43)   
 Yes250.78 (0.21–2.27)   260.86 (0.31–3.26)   
Osteolysis    280.0020.012
 No70   670.26 (0.00–1.12)   
 Yes3   53.83 (2.16–5.29)   
Bony proliferation  201< 0.001< 0.006  232< 0.001< 0.006
 No520.00 (0.00–0.28)   400.05 (0.00–0.45)   
 Yes210.88 (0.24–2.98)   321.15 (0.33–2.77)   
Periostitis  125< 0.001< 0.006  2240.0010.006
 No590.00 (0.00–0.35)   550.19 (0.00–0.97)   
 Yes141.12 (0.44–3.50)   171.12 (0.39–3.50)   
Bony ankylosis  30< 0.001< 0.006  101< 0.001< 0.006
 No670.05 (0.00–0.70)   570.19 (0.00–0.51)   
 Yes63.89 (0.89–5.34)   152.62 (1.12–4.24)   

Any association between the presence of the above radiologic features and more rapid progression of disease as measured by a change in MSS was investigated. It was assessed whether the presence of any of the additional radiologic features at baseline identified patients with an increased rate of progression. The presence of soft-tissue swelling (P < 0.001), but not periarticular OP, osteolysis, bony proliferation, or periostitis, was associated with an increased rate of change of the combined PsA-MSS. There was a weak association with presence of bony ankylosis at baseline and an increased rate of progression of MSS (P < 0.05) that became insignificant after adjusting for multiple comparisons.

DISCUSSION

The findings in this study are in keeping with other studies that have used a variety of radiologic scoring systems to show that PsA progresses beyond early disease. At the end of 1 year, Kane et al and Harrison et al each found that erosive disease was found in 22–27% of patients (8, 12). Between 5 and 10 years of disease duration, radiologic damage was found in 40–57% of patients (4). In our cohort, with a median disease duration of 12 years at followup, 74% of patients had evidence of radiologic damage. This is similar to findings by Marsal et al, who documented erosive disease in 71% of patients with a median disease duration of 15 years (19). Moreover, those with baseline damage had a greater capacity for future damage. Similar studies illustrating that once radiologic damage has occurred further damage is likely are also found in RA (20).

The relationship between the Sharp score and clinical disease in RA is an intriguing one. Pincus et al found that the Sharp score correlated well with joint deformity (r = 0.68), but poorly with joint tenderness and swelling (r = 0.01 and r = 0.19, respectively) in RA (9). A good correlation between the MSS and clinical joint scores was observed in our study, although it must be emphasized that the joint score used encompassed both synovial swelling and joint deformity. Nevertheless, it is interesting that despite the heterogeneity of PsA, involvement of the hand is an accurate reflection of global joint disease assessed clinically, lending further weight to the construct validity of the MSS. Other groups have also observed the link between erosions and clinical disease in PsA. Marsal et al found that erosions on hand and foot radiographs were correlated to the total number of joint areas ever noted to be swollen (19). Because involvement of the feet may be an important feature of PsA, we will investigate any additional value of scoring radiographs of the feet in future work, but have not done so at present.

Welsing et al have speculated that in RA patients with early disease, functional status may reflect joint inflammation, whereas joint damage has a larger impact on functional status in late disease (21). Kane et al also found that in early PsA, despite progressive radiologic damage, functional status improved (12). This was presumably secondary to better control of joint inflammation with disease-modifying antirheumatic drugs (DMARDs). In the current study, we observed a weak correlation between HAQ scores and MSS at a median of 5 years of disease duration, and only moderate correlation at 12 years. However, the relationship between functional status and its potential determinants in PsA is thought to be more complex. Sokoll and Helliwell found that despite greater peripheral joint damage in patients with RA, patients with RA and PsA had similar functional status (22). The authors speculated that other factors, possibly psychological (such as depression and anxiety), and the additional burden of skin disease might be important contributory determinants of functional status in PsA (22).

The high reliability of the MSS has been observed in other studies (14), although it must be emphasized that the numbers we tested for reliability were small and did not encompass a broad spectrum of disease. Moreover, training to achieve proficiency and consistency was time consuming. As in other scoring methods, a ceiling effect applies to the Sharp score, and any ongoing destruction in a joint with >50% damage is not reflected. There are also inevitable difficulties in interpreting damage (especially JSA) due to OA as opposed to PsA, particularly in the DIP joint.

Few studies have been published about PsA that have determined a rate of progression of the radiographic score. The median rate of progression of MSS in our study was +1.08 units/year. Our mean annualized rate of progression was +2.85 MSS units/year, compared with +1.00 PsA-MSS units/year in the placebo arm of a trial of etanercept treatment in PsA (11). Our rate is derived from an average over nearly 6 years of the radiologic progression in patients who have a median disease duration of 12 years at followup, whereas that in the etanercept trial was the rate calculated in 1 year in patients with a median disease duration of 9 years. The mean rates of progression in patients with RA scored with the MSS are higher than in patients with PsA. This may reflect more consistent and greater involvement of the hand in RA, even in patients taking DMARDs. Mean rates of progression of the MSS in early patients with RA receiving hydroxychloroquine, sulfasalazine, or methotrexate vary from 3.2.8 to 8.6 MSS units/year (23).

The strong correlation (r = 0.83–0.86) between JSA and erosion score suggests that they are a common end point in the pathogenesis of PsA. In the original application of the Sharp score in RA, there was a similarly high correlation between JSA and erosions (r = 0.79) (24). The relationships between the MSS and other radiologic features are particularly interesting. First, the presence of soft-tissue swelling at baseline was predictive of an increased rate of change of the combined PsA-MSS, and it reinforces the clinical studies that have found that correlates of disease activity such as the number of actively inflamed joints at baseline are predictive of future damage (3, 25). Second, the high correlation of certain radiologic features such as bony proliferation, periostitis, and bony ankylosis with the MSS implies close linkage between damage and the reparative process that is a particular feature of PsA. A radiographic scoring method developed by Wassenberg et al specifically for PsA that scores joints separately for destruction and proliferation also observed that these processes nearly always develop in the same direction (26). These observations, which demonstrate close linkage between the MSS and other radiologic features of PsA, add to the construct validity of the PsA-MSS scoring system.

In conclusion, the current study shows that radiologic damage, as assessed by the PsA-MSS, is progressive beyond early disease. There were also positive associations with clinical joint disease as well other radiologic features of PsA. In addition, patients with radiologic damage at baseline had a greater capacity for further damage, which reinforces the need to monitor this group closely. Further studies will need to assess the reliability of the PsA-MSS in larger numbers of patients over a greater spectrum of disease, in addition to determining the smallest detectable difference, minimal clinically significant change, and potential markers of radiologic progression. It will also be of interest to determine whether newer treatments such as anti–tumor necrosis factor agents are effective in slowing all aspects of radiologic progression in PsA.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. McHugh had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Ravindran, Balakrishnan, Jones, Korendowych, McHugh.

Acquisition of data. Ravindran, Balakrishnan, Jones, Korendowych, McHugh.

Analysis and interpretation of data. Ravindran, Cavill, Jones, Korendowych, McHugh.

ROLE OF THE STUDY SPONSOR

Abbott Laboratories had no role in the study design, data collection, data analysis, writing of the manuscript, or approval of the content of the submitted manuscript. Publication of this article was not contingent on the approval of Abbott Laboratories.

Ancillary