SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Objective

To assess currently available tools for measurement of enthesitis and to develop a new instrument specifically for use in psoriatic arthritis (PsA).

Methods

Twenty-eight patients with PsA underwent clinical assessment over a period of 6 months after change of disease-modifying therapy, usually to methotrexate. Measures of enthesitis included the Mander Enthesitis Index (MEI), the Maastricht Ankylosing Spondylitis Enthesitis Score, the Major index, and the Gladman index. Data from these assessments were used to develop a new enthesitis index, the Leeds Enthesitis Index (LEI).

Results

An iterative process of data reduction enabled derivation of the LEI index, which consisted of 6 sites: right and left Achilles insertions, medial femoral condyles, and lateral epicondyles of the humerus. All measures showed significant change from baseline but only the LEI and Gladman indices showed a large effect size. All indices demonstrated a floor effect (a score of 0 when the MEI is >0) but this was minimal for the LEI index. All indices correlated strongly with each other and other measures of disease activity.

Conclusion

The LEI is a new enthesitis index designed for use in PsA. It shows good test characteristics that suggest it will be a robust and reliable assessment tool. We suggest that it be adopted for use in any randomized controlled trials and longitudinal observation cohorts involving patients with PsA.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

The development of outcome assessment criteria in rheumatology has become vital to the validity of trials and their clinical application. Researchers have been slow to develop classification and diagnostic criteria for psoriatic arthritis (PsA) (1), and the development of validated outcome measures has been similarly retarded. Many of these difficulties are due to the diverse features of PsA that may resemble ankylosing spondylitis (AS) and rheumatoid arthritis in some cases but that are different in others.

It has been suggested by McGonagle et al that enthesitis underpins most if not all of the manifestations of PsA (2). The authors of a recent review of assessment measures for PsA recommended inclusion of an enthesitis measure in any clinical trial because enthesitis is a key component of disease expression (3). Three tools have been developed for AS: the Mander Enthesitis Index (MEI) (4), the shorter Maastricht Ankylosing Spondylitis Enthesitis Score (MASES) (5), and the Major enthesitis index (6). Furthermore, the Spondyloarthritis Research Consortium of Canada has studied, in 10 patients with PsA, an empirical index derived from sites included in the MASES; this is referred to as the Gladman index (7).

The MEI has been criticized for being time consuming and potentially causing distress to the patient (5). It may also include sites similar to those of fibromyalgia syndrome. Of note, the MEI also includes “insertions of the Achilles tendon and plantar fascia into the calcaneus,” although these sites were not commonly tender in the authors' pilot study (4).

The MASES was developed on this background of needing a measure with similar sensitivity to the MEI but with greater clinical utility. The MASES appears to have a small floor effect (not picking up cases with low MEI scores), but the significance of this is uncertain. Interestingly, Achilles tendon insertion is part of the MASES.

Gladman et al (7) assessed a variety of measurements in patients with PsA to determine the reliability and feasibility of the measurements for inclusion in clinical trials. Entheseal points at the rotator cuff insertion, the tibial tuberosity, and the posterior and inferior calcaneal sites were all palpated for tenderness. The authors found moderate agreement at 3 entheseal sites but poor agreement at the shoulder. This may have been due to difficulties with anatomic localization. They stated that “the use of a more extensive enthesitis measure that includes 13 sites is not likely to produce better agreement” (7).

In a recent study on infliximab in patients with AS, Braun et al (6) used an enthesitis index composed of 12 entheses that are reported to be commonly affected in the inflammatory process in AS (Major enthesitis index) (6). This index includes the iliac crests, the greater trochanters of the femur, the medial and lateral condyles of the humerus, and the insertion of the Achilles tendon and insertion of the plantar fascia to the calcaneus. The Major index has not been widely used or studied.

The recent published therapeutic trials of biologic agents in PsA that have included enthesitis considered only Achilles tendon insertion, plantar fascia, or both sites (8–11) and these were assessed as tender or not tender. There is need for an enthesitis index derived from patient data for PsA. The present study, an open-label, observational trial designed to assess the response to change of the MEI, MASES, Gladman index, and Major index, was also designed to develop a specific tool for enthesitis in PsA.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

The study took place in a secondary care setting in Bradford National Health Service Trust in West Yorkshire, UK. Approval was given by the local ethics committee. Patients with active disease, i.e., who were intolerant or unresponsive to their current disease-modifying therapies or had not yet begun disease-modifying therapy, and who presented with enthesitis were invited to participate as they presented to the outpatient rheumatology clinic. A full study information sheet and written consent form was provided. Patients who provided written consent to participate underwent an initial assessment of joint and skin disease. Drug therapy was chosen by the managing clinician as was considered appropriate. Patients underwent further clinical assessments at baseline (start of new drug therapy), 2 weeks, 1 month, 3 months, and 6 months. The procedures followed were more intensive than normal but the care of the patient, and the choice of drug, were no different from normal clinical care.

Clinical assessment included the following instruments: the MEI, MASES, Gladman index, Psoriasis Area and Severity Index (12), a 78-joint count (78 tender and 76 swollen), a measure of acute-phase response (C-reactive protein [CRP]), the Health Assessment Questionnaire (HAQ) (13), a patient pain visual analog scale (VAS), and a patient and physician global VAS. The VAS was assessed with both a 100-mm scale and a Likert scale (range 1–5). Together these measurements facilitate calculation of the American College of Rheumatology 20% response criteria (ACR20) (14) and the Psoriatic Arthritis Response Criteria (PsARC) (15) in addition to the Disease Activity Score in 28 joints (DAS28) (16). Further validation of the ACR20 and DAS28 in PsA is required, but preliminary data suggest that they may be valid in polyarticular PsA (17). The clinical assessment was performed by a single observer. Pain using the MEI was scored as follows: 0 = no pain, 1 = pain noted, 2 = wince, 3 = wince and withdraw.

Statistical analysis

The data were analyzed with SPSS 12 (SPSS, Chicago, IL). Descriptive statistics are mean values. Responsiveness was studied using the Cohen's d effect size statistic (18) given by difference in scores (time 1 to time 2) divided by the pooled standard deviation of the original time 1 and time 2 scores. In general, values >0.2–0.5, >0.5–0.8, and ≥0.8 indicate small, medium, and large effect sizes, respectively.

To develop a new index, a process of data reduction was performed. The iterative process used the whole data set from each time point of the study, which, allowing for missing data, gave a total of 115 assessments. As a first step, the original gradation of pain for each enthesis on the MEI was recoded dichotomously as “no pain” or “painful.” An original score of 0 was regarded as no pain; original scores ranging from 1 to 3 were regarded as painful. Only data from patients with an MEI score >0 were used for developing modifications of the MEI. Frequency tables were used to determine which specific enthesis was scored as painful most frequently. This enthesis was noted and all patients reporting this enthesis as painful were not taken into consideration in the next step. In the remaining patients, a similar analysis using frequency tables was performed to determine which enthesis was most frequently scored as painful. Again the patients reporting this particular enthesis as painful were not included in subsequent steps. This process was repeated until up to 80% (as in the MASES development) of patients in all data sets with an MEI score >0 were included. If an equal amount of patients reported 2 different entheses as most frequently painful at the same analysis, we selected the point that was anatomically the most easy to localize. If an enthesis was situated on the right or left side of the body, the contralateral enthesis was also included. This PsA-specific index was called the Leeds Enthesitis Index (LEI). These points are detailed in Table 1.

Table 1. Leeds Enthesitis Index (LEI) examination points
LEI examination points
Lateral epicondyle, left and right
Medial femoral condyle, left and right
Achilles tendon insertion, left and right

Spearman's correlation was used to determine the relationship between the entheseal indices and disease activity as measured by the swollen and tender joint counts, patient and physician global VAS, CRP level, HAQ, and DAS28. To assess a possible floor effect, the number of patients with an enthesitis index score of 0 and a score on the MEI >0 were identified.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

The baseline characteristics of the cohort are described in Table 2. The HAQ, tender joint count, swollen joint count, physician and patient global VAS, and DAS28 indicated active disease with significant impact. Methotrexate was initiated in 19 patients, leflunomide in 4, etanercept in 4, and hydroxychloroquine in 1. Patients showed a good response to treatment with acceptable PsARC and ACR20 responses at both 3 and 6 months. The DAS28 indicated a good response at 3 months and moderate response at 6 months. These results are shown in Table 3.

Table 2. Baseline characteristics*
CharacteristicResult
  • *

    Values are the mean ± SD (range) unless otherwise indicated. CRP = C-reactive protein; HAQ = Health Assessment Questionnaire; VAS = visual analog scale; DAS28 = Disease Activity Score in 28 joints.

Sex, male:female14:14
Age, years46.5 ± 10.5 (24–70)
Disease duration, years10.5 ± 11.3 (0.8–37)
CRP, mg/liter (normal <8)15.3 ± 13 (5–51)
HAQ1.4 ± 0.8 (0–2.9)
Tender joint count16.3 ± 12.5 (1–47)
Swollen joint count8.9 ± 4.3 (1–20)
Physician global VAS51.3 ± 21.1 (15–89)
Patient global VAS56.2 ± 20.6 (23–100)
DAS284.3 ± 1.0 (2.2–6.7)
Table 3. Disease response measures*
Disease response measuresBaseline3 months6 months
  • *

    PsARC = Psoriatic Arthritis Response Criteria; ACR20 = American College of Rheumatology 20% response criteria; DAS28 = Disease Activity Score in 28 joints.

  • P < 0.01.

PsARC, %5461
ACR20, %5043
DAS28 score, mean4.33.03.4

The iterative reduction process took 3 rounds to identify 80% of our patients. The first round produced the right lateral epicondyle (49%). The second round produced the right medial femoral condyle (70%). The third round produced 3 sites: right posterior superior iliac spine, cervical spinous processes, and left Achilles tendon insertion (80%). The Achilles tendon was chosen because it was the easiest site to localize. The contralateral point for each site was added, bringing the total to 6. Tenderness at these 6 points was detected in 80% of patients with an MEI score >0, suggesting that a positive score with this combination is likely to be sensitive enough to monitor entheses that are tender.

The mean scores of the 5 enthesitis instruments at baseline, 3 months, and 6 months are indicated in Table 4. The scores showed a steady improvement over the study period. However, the scores for the MEI, MASES, and Gladman index were lower at 3 months than at 6 months whereas the LEI and Major scores improved steadily in line with the clinical indices.

Table 4. Outcomes for enthesitis indices*
 Baseline3 months6 months6-month effect size
  • *

    Values are the mean ± SD unless otherwise indicated. LEI = Leeds Enthesitis Index; MEI = Mander Enthesitis Index; MASES = Maastricht Ankylosing Spondylitis Enthesitis Score.

  • Effect size <0.2 = nil, 0.2–0.5 = small, >0.5–0.8 = moderate, >0.8 = large.

  • P < 0.05, Wilcoxon's signed rank test.

  • §

    P < 0.01, Wilcoxon's signed rank test.

LEI2.9 ± 1.71.9 ± 1.71.6 ± 1.7§0.82
MEI11.6 ± 7.67.2 ± 6.57.6 ± 8.90.50
MASES4.9 ± 3.92.2 ± 2.1§3.1 ± 3.80.49
Gladman index2.5 ± 1.60.6 ± 0.8§0.8 ± 1.3§1.22
Major index3.9 ± 2.22.6 ± 2.1§2.1 ± 2.4§0.80

The 5 measures were analyzed for response to change and whether there was any significant difference at visits 4 (3 months) and 5 (6 months). These results are shown in Table 5. The effect size at 6 months was large for the LEI and Gladman index, moderate for the Major index, and small for the MASES and MEI. The indices from the tumor necrosis factor trials (Achilles and Achilles/plantar fascia) were unresponsive at 3 months and only the Achilles/plantar fascia was significantly different at 6 months, with a small effect size (data not shown).

Table 5. Analysis of discrimination*
 Baseline (n = 28)3 months (n = 24)6 months (n = 26)
  • *

    Values are the number of patients with an enthesitis score of 0 when Disease Activity Score in 28 joints is >2.8. See Table 4 for definitions.

LEI233
MEI111
MASES946
Gladman index969
Major index124

To determine if the indices were good at detecting active disease, we used a DAS28 cutoff of 2.8. The LEI and MEI were best able to distinguish between patients with active disease and those without (Table 5). The number of patients with an index score of 0 who also had a DAS28 score >2.8 was smallest for the LEI, Major index, and MEI.

When looking for a floor effect, we checked how many patients scoring 0 on any of the indices also had an MEI score >0. The LEI had fewer cases in this category at each time point (Table 6). Achilles and Achilles/plantar fascia showed poor discrimination and a significant floor effect (data not shown).

Table 6. Analysis of floor effect*
 Baseline (n = 28)3 months (n = 24)6 months (n = 26)
  • *

    Values are the number of patients with an enthesitis score of 0 when Mander Enthesitis Index score is >0. See Table 4 for definitions.

LEI123
MASES845
Gladman index7810
Major index334

The correlation matrix between entheseal indices and clinical activity scores is shown in Table 7. All of the enthesitis indices correlated strongly with each other at each time point (data not shown). There was an association between all indices and the physician global VAS, patient global VAS, and DAS28. Tender joint count related strongly to the LEI, MEI, MASES, and Gladman index, while the swollen joint count related to the LEI and Major index only. The HAQ and patient pain VAS correlated with all indices except the Gladman index. Only the MASES showed a relationship with CRP. The clinical parameters correlated most consistently with the LEI.

Table 7. Spearman's correlation matrix between enthesitis indices and clinical measures*
 CRPPhysician global VASTJCSJCHAQPatient pain VASPatient global VASDAS28
  • *

    CRP = C-reactive protein; VAS = visual analog scale; TJC = tender joint count; SJC = swollen joint count; HAQ = Health Assessment Questionnaire; DAS28 = Disease Activity Score in 28 joints; see Table 4 for additional definitions.

  • P < 0.01.

  • P < 0.05.

LEI−0.1570.3830.4620.1970.2240.3600.3430.453
MEI−0.1260.4610.5370.1660.3810.5280.4980.541
MASES−0.2900.3660.3770.0960.3560.4290.4430.344
Gladman index−0.2040.4270.4890.3380.2440.2350.4600.651
Major index−0.0410.4630.0920.2170.2200.4130.4280.390

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Enthesitis is considered an important feature of PsA. The recent Outcome Measures in Rheumatology Clinical Trials VIII (OMERACT VIII) (19) meeting recommended that enthesitis be included in clinical studies of PsA. To be used as an outcome measure in a trial, a tool needs to fulfill a number of requirements best encapsulated by the OMERACT filter domains of truth (validity), discrimination (reproducibility and responsiveness), and feasibility.

Measuring enthesitis with a clinical tool gives us information about tenderness at a given point but does not explain what may be causing the pain. To have validity, the measure needs histopathologic or imaging confirmation. There are no strict definitions or criteria for histopathologic diagnosis of enthesitis, mainly due to the ethical and technical difficulty of obtaining pathologic material. The available biopsy studies are mainly from patients with AS or patients with a more general spondylarthritis (SpA) classification. The classic description was provided by Ball in a study of elderly patients with AS, in whom the pathologic process seems to involve the subchondral bone (20). This is the site of an erosive inflammatory infiltrate composed of lymphocytes and plasma cells and occasionally polymorphonuclear cells. The erosive process is followed by fibrous tissue proliferation leading to the formation of cartilage and, subsequently, bone. A more recent immunohistochemistry study of entheses in elderly patients with SpA undergoing total knee joint replacement demonstrated edema with an inflammatory infiltrate in the bone marrow underlying an enthesis (21). A small study of patients with an early diagnosis of SpA who had entheseal symptoms demonstrated increased vascularity and cellular infiltrate throughout the fibrous part of the enthesis. Bone and subentheseal edema were found in 3 of 5 patients, but the amount of bone tissue available for analysis was small (22). The histopathology suggests both a superficial and deeper bone component to entheseal disease that contributes to some of the difficulties in developing robust clinical outcome measures.

There have been several ultrasound studies comparing clinical detection of enthesitis in SpA with imaging confirmation. Lehtinen et al (23) examined the clinical utility of examination compared with ultrasound in the pelvis and lower limb in 31 patients. Ultrasound found abnormalities in only 33 of 56 clinically tender sites whereas tenderness was found in only 21 of 44 sites with ultrasound abnormalities. Sensitivity of the clinical examination was 37.5% and specificity was 92.7%. The closest correlation was at the plantar fascia insertion. Balint et al (24) examined 5 sites in the lower limb in 35 patients. They found an overall sensitivity of 22.6% and a specificity of 79.7%. The results may have been influenced by the use of enthesophytes (or spurs) as an indicator of enthesitis. Because enthesophytes are frequently present in the normal population, they may have decreased specificity. D'Agostino et al (25) used power Doppler ultrasound to examine 164 patients with SpA; of these, 60 had PsA. Only 34 of these patients were examined clinically. The authors found a sensitivity of 41% and specificity of 65%. The patients with PsA had abnormal entheses on ultrasound at 128 of 378 sites examined. However, these studies may not be comparable because each used a different sonographic definition of enthesitis, but they do suggest that it is difficult to be confident that a tender point represents an inflamed enthesis when evaluated clinically and casts some doubt on the validity of tender points to indicate enthesitis.

Although ultrasound is an excellent tool, it may miss some important deeper bony changes. Magnetic resonance imaging (MRI) detects bone edema and osteitis commonly associated with entheseal disease, and this modality may provide better correlation with clinical findings. Marzo-Ortega et al (26) reported clinical and MRI improvement of symptomatic entheseal sites in the spine in 4 patients from a larger cohort with SpA. Unfortunately, similar correlation with peripheral entheseal sites was not available, although in a separate study clinically apparent enthesitis around the heel was associated with widespread osteitis on MRI (27).

A further area of uncertainty concerns the juxtaarticular site of some entheseal points. It may be that synovitis causes diffuse pain and this will result in a positive enthesitis point. We believe that the level of synovitis likely to cause diffuse pain at the joints near the LEI sites (elbow, knee, and ankle) could be easily detected on clinical examination: in the current series, none of the patients had significant juxtaarticular disease likely to cause this confounding (data not shown).

The MEI, LEI, and Major index were best able to distinguish between patients with active disease and those without active disease as defined by the DAS28. This finding suggests that these measures relate to an assessment of inflammation on a more general scale and probably do represent localized inflammation at the entheseal points. The floor effect was lowest for the LEI, meaning that it will be likely to detect the large majority of patients with active enthesitis.

Heuft-Dorenbosch et al (5) questioned whether a low MEI score represented clinically important enthesitis and commented that the improvement in feasibility of their new index (MASES) would more than balance out the nondetection of a few patients. They noted a moderate floor effect with both the MASES and Major indices. The data show that the LEI and Major index both have a floor effect, with the LEI having the least floor effect. Again, the large gain in feasibility could be expected to outweigh the nondetection of a few patients.

The reliability of the LEI, MASES, and Major index was assessed during the International Spondyloarthropathy Interobserver Reliability Exercise study (28). The intraclass correlation coefficients for the LEI, MASES, and Major index were 0.81, 0.56, and 0.7, respectively, suggesting that they are reliable and repeatable tests.

The LEI, Gladman index, and Major index showed a large effect size at 6 months and significant response to change, while the MEI and MASES showed a small effect size but a significant change from baseline. The correlation results suggest that the associations are most robust for the LEI. These results together suggest that the LEI is at least as good as any of the other indices and probably superior in some respects.

The ease of use often determines if a measure is adopted or not. If a measure is time consuming or requires extra resources, it will not be used no matter how truthful or discriminatory it is. The LEI considers only 6 sites and therefore takes little time. The sites are also easy to identify anatomically, suggesting a minimum amount of training will be needed for metrologists involved in collecting data for trials or clinics. While the LEI is the quickest and easiest, the other measures, except the MEI, are also simple to perform.

The LEI is a new enthesitis index designed for use in PsA. It shows good test characteristics that suggest it will be a robust and reliable assessment tool. We suggest that it be adopted for use in any randomized controlled trials and longitudinal observation cohorts involving patients with PsA.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Mr. Healy 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 design. Healy, Helliwell.

Acquisition of data. Healy.

Analysis and interpretation of data. Healy, Helliwell.

Manuscript preparation. Healy, Helliwell.

Statistical analysis. Healy, Helliwell.

ROLE OF THE STUDY SPONSOR

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

This was an investigator-initiated study partly funded by Sanofi-Aventis. Sanofi-Aventis had no role in the trial design, data collection, analysis, or writing of the manuscript.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES