PATIENTS AND METHODS

Patients

Patients satisfying the European Spondylarthropathy Study Group criteria for the diagnosis of SpA²³ were assessed during routine presentation to the rheumatology outpatient clinic. Patients with previous joint surgery of the knee and ankle, corticosteroid injection of the structures examined within the previous six weeks, or peripheral neuropathy of the lower limbs were excluded from the study.

Clinical examination

The superior pole of the patella (quadriceps tendon insertion), the inferior pole of the patella (patellar ligament origin), the patellar ligament insertion at the tibial tuberosity, the Achilles tendon, and the plantar aponeurosis were examined in both lower limbs of each patient. Clinical examination for tenderness and swelling at each site was performed by an experienced rheumatologist.

Ultrasound evaluation

Real time ultrasonography was performed by an experienced rheumatologist trained in musculoskeletal ultrasonography, using an ATL HDI 3000 machine with linear 7–4 MHz and compact linear 10–5 MHz probes. The linear 7–4 MHz probe was used for detecting suprapatellar bursitis as it allows deeper penetration when a large effusion is present and was preferred for imaging of the plantar aponeurosis. The clinical examination and US measurements were performed separately; one immediately after the other, by different investigators who were unaware of each other's results. The sonographer did not know the clinical details of the patients. Examination of the superior pole of the patella (quadriceps tendon insertion), the inferior pole of the patella (patellar ligament origin), and the patellar ligament insertion at the tibial tuberosity was performed with the patient in the supine position with the knee flexed at 30 degrees. The Achilles tendon and the plantar aponeurosis were examined with the patient lying prone with the feet hanging over the edge of the examination table at 90 degrees of flexion. If the patient was unable to lie prone, the heels were examined with the patient lying supine and the knees and ankles flexed at 90 degrees.

Ultrasonographic assessment of structure thickness and the presence or absence of bony erosion, enthesophyte, and bursitis was recorded at each site. Bursitis was defined as a well circumscribed, localised anechoic or hypoechoic area at the site of an anatomical bursa and which was compressible by the transducer. Bursal dimensions were obtained in the long and short axis with a normal bursa being <2 mm in the short axis.²⁴ Bony erosion was defined as a cortical breakage with a step down contour defect and an enthesophyte was defined as a step up bony prominence at the end of the normal bone contour. Ligament, aponeurosis, and tendon thickness were measured at the point of maximal thickness proximal to the bony insertion. The following criteria were used for abnormal structure thickness: quadriceps tendon thickness ≥6.1 mm,¹⁹ proximal and distal patellar ligament ≥4 mm,²⁰ Achilles tendon ≥5.29 mm,²¹ plantar aponeurosis ≥4.4 mm.²² In this study only thickened enthesis, fluid collection, erosion, and bony spur were accepted as US signs of enthesitis. “Hypoechoic oedema” without any thickness change was excluded, as it is a subjective sign of enthesitis. No control group was selected in this study as the normal ultrasound features and dimensions of the structures examined have already been described.^19–22 To reduce subjectivity further the threshold of abnormal thickness was set 0.1 mm above the reported standard deviation of each site in the normal population.

An ultrasonographic score of lower limb enthesitis was calculated as follows: one point was scored for each abnormality at each site examined, giving a possible maximum total score of 36 (box 1). The US score was also calculated separately as a soft tissue score and as a bone score. The soft tissue score included entheseal thickness and bursitis. The bone score included erosions and enthesophytes. Interobserver error was not measured in this study, but we have already demonstrated a high degree of intraobserver repeatability and interobserver reproducibility of US measurements of ligament thickness.²⁵ Sonographic images were stored on magneto-optical disks for offline analysis. After three months, the investigator rescored every patient to calculate intraobserver reliability.

Statistical analysis

All values are given as mean (median) (standard deviation). Statistical analysis was performed using Statview software. A p value <0.05 was deemed significant. Intraobserver agreement was calculated using a κ test. Sensitivity, specificity, negative and positive predictive value, false negative and false positive rates of clinical examination were also calculated.

RESULTS

Patient characteristics

Thirty five patients (25 male, 10 female) were examined (ankylosing spondylitis 27; psoriatic arthritis 7; reactive arthritis 1). The patients had a mean age of 48 (median 49.3) (SD 14), mean age of disease onset of 20.8 (median 19) (SD 13.1) years and a mean duration of disease of 24.9 (median 21) (SD 10.4) years. The mean erythrocyte sedimentation rate (ESR) was 16.2 (median 10) (SD 18) mm/1st h and the mean C reactive protein (CRP) was 18 (median 6) (SD 20.8) mg/l. Six (17%) patients had a family history of spondyloarthropathy, 14 (40%) had previous uveitis, 21 (60%) had peripheral joint disease, one (3%) had inflammatory bowel disease, nine (26%) had psoriasis, and two (6%) had joint replacement surgery not affecting the lower limbs. Twenty four (69%) patients were HLA-B27 positive, three (9%) were negative, and the HLA status of eight (23%) was not known.

Clinical examination

A total of 350 entheseal sites were examined (10 sites×35 patients) clinically, of which, 71/350 (20%) entheseal sites were tender and 13/350 (4%) entheseal sites were swollen. Table 1 gives the clinical findings at individual sites. On clinical examination 61.7% of entheses were symmetrically involved.

Ultrasonographic examination

A total of 350 entheseal sites were examined by US. Two sites were obscured by overlying psoriatic plaques (one over the inferior pole of the patella and one over the tibial tuberosity in two separate patients) and are not included in the analysis. Table 1 gives the ultrasound findings at individual sites. On US examination 195/348 (56%) entheses were abnormal and 54% of the entheses were symmetrically involved. Figure 1 shows the ultrasonographic features of the five normal entheseal sites and fig 2 the ultrasonographic features of enthesitis. The intraobserver κ value for analysis of all sites was 0.9. κ Values for analysis of the separate lesions were (bony erosion κ=0.97, bursitis κ=1.00, enthesophyte κ=0.83) and for analysis of separated locations were (superior pole of the patella κ=0.92, inferior pole of the patella κ=0.7, tibial tuberosity κ=0.82, superior pole of the calcaneus κ=0.97, inferior pole of the calcaneus κ=0.8).

Relationship of clinical and ultrasonographic findings in the lower limb entheses

Entheseal abnormality was detected by US in 195/348 (56%) entheseal sites and by clinical examination in 75/348 (22%) sites. Figure 3 shows the relationship between clinical and ultrasonographic findings at individual sites.

Table 2 gives the sensitivity, specificity, negative and positive predictive value, false negative and false positive rate of clinical examination. Compared with US, clinical examination had a low sensitivity (22.6%) and moderate specificity (79.7%) for the detection of enthesitis of the lower limbs.

Ultrasonographic score of enthesitis

To quantify lower limb enthesitis a US score of enthesitis was formulated. A maximum of 36 was possible in each patient. The mean score was 6.9 (median 6) (SD 4.3) with a soft tissue US abnormality score of 5 (median 4) (SD 2.8) and a bone (erosion or/and enthesophyte) US score of 1.9 (median 2) (SD 1.8). There was no significant correlation between the US score of enthesitis and the ESR or CRP.

Clinically detected and undetected bursitis

There was no correlation between the sizes of the bursae on US and clinical findings and only 5/19 bursa were clinically detected. Two large bursae were detected clinically but two equally large bursae were not detected by clinical examination. No bursa under 13.8×4.7 mm was detected on clinical examination.

DISCUSSION

On clinical examination, 75/348 (22%) entheseal insertions were determined to be inflamed, which is similar to a previous study of calcaneal entheses.²⁶ Lehtinen et al noted less clinical enthesitis in lower limbs in SpA (56/372 (15%)),¹⁰ which may reflect differences between the two patient groups. The presence of tenderness was more sensitive than swelling in the detection of enthesitis, being present at 14–28% of entheseal sites while swelling was present at 0–14% of entheseal sites. US detected enthesitis at 51.4–63.8% of entheseal sites and was more sensitive than clinical examination for tenderness and swelling taken together or separately. If US is taken as the gold standard, neither tenderness nor swelling was specific in the detection of enthesitis, with just 44/75 clinically inflamed entheses confirmed by US. Some of these sites might have had minor degrees of inflammation not amenable to detection on US. Possibly, certain features of enthesopathy such as plantar spur were overestimated as they may be found in normal subjects, though the presence of erosion and bursitis at the enthesis is relatively specific. In this study, swelling was not a useful sign in detecting enthesitis at the tibial tuberosity or plantar aponeurosis.

Two principal features of soft tissue inflammation—tendon thickening and bursitis—were examined on ultrasonography. To reduce the subjectivity of previous studies of enthesitis in SpA, tendon thickness was determined using standardised US views according to previously published protocols. In addition, the threshold of normal tendon thickness was set at 0.1 mm above the reported standard deviation above the mean as measured in normal populations. We have already demonstrated a high degree of intraobserver repeatability and interobserver reproducibility of US measurements of ligament thickness.²⁵ Increased tendon or ligament thickness was the most common feature of soft tissue inflammation, being present at 20–57% of entheses examined. This study may underestimate enthesitis as it did not include hypoechoic changes in tendons or ligaments, which are an important but subjective feature of enthesitis and may be influenced by transducer position,²⁷ particularly when parallel fibrils change their directions as is the case at the entheses. Tendon thickening may be due to either oedema or fibrosis, and longitudinal studies are required to assess its reversibility. Significant qualitative and quantitative ultrasonographic changes were demonstrated in a case study of Achilles tendonitis and retrocalcaneal bursitis followed up for one year,²⁸ whereas Lehtinen et al found no reduction in the frequency of ultrasonographic enthesitis in a six month follow up of 23 patients.¹⁶ Bursitis was present at 4–13% of sites examined, being most common at the suprapatellar and retrocalcaneal sites.

Enthesophyte formation was the most common US bony abnormality, being present at 3–30% of entheseal sites while bony erosion was present at 1–13% of sites. This is in keeping with previous radiological studies of SpA where entheseal ossification is a common feature. In addition to entheseal ossification, US also demonstrated intratendinous and interligamentous calcifications adjacent to the point of entheseal insertion. These may represent the end stage of inflammation or may relate to other pathology such as trauma or degenerative changes which are common in the general population. Erosive changes at the enthesis may have been underestimated owing to the presence of an enthesophyte which obscures adjacent erosions on US. Clinical examination is not useful in determining the presence of these bony lesions and does not provide this potentially diagnostic information.

This study confirms that there is considerable subclinical enthesitis in SpA, which can be objectively measured by US using standardised protocols. At present, only clinical evidence of enthesitis of the heels is included in the European Spondylarthropathy Study Group preliminary classification criteria for the diagnosis of SpA,²³ and clinical enthesitis is included in the preliminary core sets to be used as end points in clinical trials in ankylosing spondylitis.²⁹ US detection of enthesitis is more sensitive and more specific than clinical examination and soft tissue US has a high degree of reproducibility. Ultrasonography is now widely practised by rheumatologists and should be used to define classification criteria and outcome measures in SpA.

Although established treatments of SpA have not been proved to reduce spinal entheseal manifestations of disease,³⁰ newer treatments such as anti-tumour necrosis factor have been shown to reduce spinal entheseal inflammation.^31,32 Studies of treatments specifically targeting entheseal inflammation will need objective measures of peripheral enthesitis to confirm efficacy. Plain radiography is limited in that established scores¹² provide little information about soft tissue inflammation, which is most amenable to treatment. MRI and US have similar sensitivity in demonstrating superficial soft tissue and bony surface abnormality.^14,15 US does not detect insertional bone oedema, but the significance of this lesion is not yet determined in SpA and bone oedema may also co-exist in overuse syndrome with enthesopathy.³³

By using a limited series of easily reproducible, fixed reference points and established normal parameters we devised an enthesitis score of the lower limbs: GUESS, which can be applied to the evaluation of treatments in SpA. US did not correlate with systemic parameters of disease activity in SpA (as it does in rheumatoid arthritis³⁴), though these are less sensitive markers of disease activity in SpA.³⁵ Ultrasonographic assessment is more time consuming than clinical examination but can be reliably performed in 15 minutes by an experienced musculoskeletal ultrasonographer. Further multicentre studies of GUESS in the assessment of enthesitis are required.