Home

Current Issue

Archives

Guidelines for Authors

Classified Ads

Links

Search PubMed

Subscriptions

Subscriber Registration

Guidelines for Website Users

JRheum Update Service

Contact Info

JOHN C. DAVIS Jr, MD, MPH,
Associate Professor of Medicine,
Director, Clinical Trials Center,
Division of Rheumatology,
University of California,
San Francisco, California, USA;
DAFNA D. GLADMAN, MD, FRCPC,
Professor of Medicine,
University of Toronto,
Senior Scientist, Toronto Western Research Institute,
Deputy Director, Centre for Prognosis Studies in the Rheumatic Diseases,
Toronto, Ontario, Canada

Address reprint requests to Dr. D. Gladman, Toronto Western Hospital, 1E-410B, 399 Bathurst Street, Toronto, Ontario M5T 2S8. E-mail: dafna.gladman@utoronto.ca

Disease activity in ankylosing spondylitis (AS) is measured by multiple measures including metrology, radiology, laboratory, functioning, and disability¹. Spinal mobility assessment is advantageous, as loss of mobility can be an early feature and is utilized for classification criteria². Loss of spinal mobility has also been reported to be a prognostic factor³. Significant therapeutic advances have changed the way that patients with AS are treated and have shown the potential for disease modification⁴. The Assessment in Ankylosing Spondylitis (ASAS) group has recommended that spinal mobility measures be used as part of the disease controlling antirheumatic therapy criteria⁵.

Over the past 50 years, multiple spinal mobility measures and composite scoring systems have been developed and are available for clinical application. Frequently these measurements have not been standardized or assessed for reliability, validity, or sensitivity to change⁶. Moreover, performance of some measures is time-consuming and may not be feasible in clinical practice. Therefore the goal would be to identify a minimum number of individual clinically appropriate measures to accurately assess the axial manifestation of disease. The purpose of this review is to assess the measures that have been used to assess spinal disease in terms of their validity and conformity with the OMERACT filter.

OMERACT FILTER

The OMERACT (Outcome Measures in Rheumatoid Arthritis Clinical Trials) process involves consensus on outcome measures and is based on the "OMERACT filter," composed of 3 key components: truth, discrimination, and feasibility⁷. Truth addresses the issues of face, content, construct, and criterion validity -- does the instrument measure what it is intended to? Is the result unbiased and relevant? Discrimination addresses issues of reliability (inter and intraobserver) and sensitivity to change. Feasibility addresses whether the measure can be easily applied or used with regard to time, money, training, scoring, interpretability, and acceptance by physician and patient.

Table 1 lists the various areas that have been assessed. This review deals only with those measures that have been tested and reported in the literature (Table 2).

Table 1. Types of metrology measures described/studied.

Table 2. Reliability, sensitivity to change and feasibility in instruments used to assess spinal disease in ankylosing spondylitis.

CERVICAL SPINE

Cervical spine movement includes lateral flexion, rotation, and forward flexion.

Lateral flexion of the neck using an inclinometer was found to correlate moderately with radiological change (r = 0.57), with excellent inter- and intraobserver reliability [intraclass correlation coefficient (ICC) 0.90 and 0.96, respectively]⁸. A decrease in distance between the tragus and the coronoideus process of the clavicle in maximal lateral bending of the head to the left and right measured by a tape was found to correlate better with radiological change (r = 0.60) and had excellent intra- and interobserver reliability (ICC 0.97 and 0.98, respectively)⁸.

Cervical rotation has been tested in several ways. Using a Myrin inclinometer, full rotation from left to right was measured⁸, with moderate correlation with radiological change (r = 0.41), and with excellent inter- and intraobserver reliability (ICC 0.98 and 0.96, respectively). With a tape measure, the change in distance between the chin and the coronoideus process in maximal cervical rotation from left to right has been found to correlate better with radiological change (r = 0.57) and have excellent inter- and intraobserver reliability (ICC 0.97 and 0.98, respectively)⁸. Another method using a tape measure assesses the difference in the distance between a mark in the suprasternal notch and the tragus of the right ear when the neck is rotated from left to right⁹. This method provided excellent intraobserver (ICC 0.80 and 0.89) and interobserver reliability (ICC 0.82). The Bath Ankylosing Spondylitis Metrology Index (BASMI)¹⁰ measures cervical rotation using a gravity-action goniometer. The mean of the right and left results is calculated and the results are scored as 0 if > 70°, 1 if 20°–70°, and 2 if < 20°. Interobserver reliability was 0.98 and intraobserver reliability 0.99. Thus, cervical rotation reflects changes in axial disease in AS; however, measurements that require specific instruments may be difficult to perform.

Chin to chest: This measure of forward flexion of the neck was found to correlate poorly with radiological disease, thus its relevance to the disease process (truth) is questionable. It was, however, found to be reliable, with inter- and intraobserver agreement measured by ICC of 0.92 and 0.95, respectively⁸. It has not been included as an outcome measure in clinical trials, so sensitivity to change is not available.

Occiput to wall: The distance measured between the occiput and the wall when the patient stands with heels and shoulder against the wall with the back straight has been used to assess patients for many years and is thought to correlate with kyphosis. It was correlated with radiographic change in the study by Viitanen, et al⁸ (r = 0.49). Interobserver reliability was high (r = 0.92). It also proved to be sensitive to change, with an effect size of 0.25.

Tragus to wall: The distance between the tragus of the ear and the wall is measured when the patient stands with heels and buttocks touching the wall, knees straight, shoulders back, and places the head as far back as possible, keeping the chin in. Correlation with radiographs was high (r = 0.92), and interobserver and intraobserver reliability both were very high (r = 0.99)¹⁰.

Occiput to wall and tragus to wall were studied in a longitudinal observational cohort, and both were comparably reliable¹¹. The authors recommended using the occiput to wall because a value of zero clearly distinguishes between patients with normal thoracic spine extension and kyphotic patients.

Lumbar mobility. Forward flexion.

Schober's test: In the original report in 1937¹² the subject stood upright and lumbosacral junction was identified by a skin mark. Another mark was made 10 cm above this and the distance between the 2 was recorded when the patient bent maximally forward. It was thought to reflect the movement in the lumbar spine, but criticism has been that areas of spine used are more susceptible to superficial stretching of the skin and not actual underlying structures, particularly since there was no anatomical definition of the first mark. No studies of reliability or validity were carried out initially. A study by Macrae and Wright¹³ tested the original Schober test against their modification (which added an additional mark 5 cm below the first mark), and showed that both the original Schober and their modification reflected lumbar forward flexion (confirmed by radiographs). Their modified Schober correlated with radiographic changes better than the original test (r = 0.97 vs 0.90, respectively)¹³. Testing the original Schober method, a misplacement of 2 cm at the level of the lumbosacral junction against radiographs resulted in an error of up to 15°, whereas the modified Schober method reduced the error to less than 5°.

Modified Schober test: Moll and Wright further refined the Macrae and Wright modification of the Schober test by providing an anchor for the lumbosacral junction using a line across the upper limit of the dimples of Venus¹⁴. The subject was then asked to bend maximally forward, and the new distance between the upper and lower marks was measured. The distraction of this mark has been found to correlate very closely with anterior flexion measured radiologically (r = 0.97)¹³. Moll and Wright demonstrated a moderate correlation between spinal mobility and duration of symptoms in patients with AS and no correlation with age¹⁵.

The modified Schober is included as one of the measures of the BASMI. However, in the BASMI the first mark is made as a line across the iliac crests, which is at the level of L4 rather than at the lumbosacral junction, which is marked by a line across the dimples of Venus. Unfortunately, the method was not revalidated against radiographs as the original and its initial modification have been. Within the BASMI this modification of the Schober test had criterion validity against 20 other items included in the index (r = 0.92). Interobserver as well as intraobserver reliability was demonstrated (r = 0.96 and 0.99, respectively). Viitanen, et al¹⁶ used the Macrae and Wright modification of the Schober test and found that it provided excellent inter- and intrarater reliability (ICC 0.96 and 0.94, respectively), and correlated highly with the radiographic changes in the spine (ICC 0.71). They also demonstrated the modified Schober to be sensitive to change following intensive physiotherapy (effect size 0.24). However, another study showed that the Schober test did not change following this therapy¹⁷. Moreover, although infliximab was proven to be effective in AS, the Schober test did not distinguish between patients treated with active drug or with placebo¹⁸. It should be noted that the Schober test used in this trial was based on the BASMI method.

Feasibility: The Schober test and its modification are easy to perform, and scores are easy to record and interpret, with no special equipment required; minimal training is required and patient and physician acceptance is good. Thus, this test passes the OMERACT filter of truth, discrimination (only for the original modification), and feasibility.

Lumbar flexion. Forward flexion of the spine using finger to floor distance was included among the measures studied by Viitanen, et al⁸. They found that it had good reliability and sensitivity, but did not correlate with radiographic changes. Heikkilä, et al¹⁷ demonstrated that finger to floor distance improved following physiotherapy in patients with AS.

Lumbar flexion/extension. A method to simultaneously assess the range of flexion and extension of the spine was described by Miller, et al¹⁹. This test uses a line across the dimples of Venus as the first landmark, from which three 10-cm segments are marked with the back fully flexed. The change in the distance between these points is noted with the back fully extended. It was compared with finger to floor distance, modified Schober test, and a goniometric method, showing excellent correlation and low interobserver error. This was the most sensitive method for detecting loss of spinal mobility, with the upper segment being most sensitive. Thus, it may be more relevant to patients with early spondylitis than the Schober test.

Lateral bending. Several methods have been developed to measure lateral bending of the lumbar spine. Moll and Wright used a measurement of the change in the distance of 2 marks inked on the skin of the lateral trunk between upright position and lateral flexion. The upper mark was placed at the point where a horizontal line through the xiphisternum crossed the coronal line, and the second mark where a horizontal line through the highest point of the iliac crest crossed the coronal line¹⁵. Although the authors demonstrated differences between patients with AS and normal controls, the method was felt to be cumbersome. Subsequent methods used the distance between the tip of the third finger and the floor when the patient stands upright, heels, buttocks, and shoulders against the wall, and bends sideways without lifting the opposite foot off the ground. This has been measured either by tape measure secured on the wall, or by making a mark on the thigh and leg^10,16. The first method has been included in the BASMI¹⁰, and the latter method in the Edmonton AS Metrology Index (EDASMI)⁹. Both methods provided excellent reliability. Both methods are relatively easy to perform (requiring the observer to bend to the ground several times) using only a tape measure and require minimal training.

Chest expansion. This is measured by the difference in chest circumference between full expiration and inspiration at the fourth intercostal space^8,16. Chest expansion did not correlate with radiographic changes (r = 0.38). However, chest expansion provided interobserver as well as intraobserver reliability (ICC 0.85 and 0.95, respectively). Sensitivity to change was noted for chest expansion, with an effect size of 0.42. Although chest expansion did not correlate with radiographic change, it did provide reliability and discrimination. This measure is included in the BASMI and the ASAS response criteria.

Intermalleolar distance. Intermalleolar distance measures abduction of the hips. Keeping the knees straight and the legs in contact with the resting surface the patient is asked to take the legs as far apart as possible, and the distance between the medial malleoli is measured. This measure had criterion validity as part of the BASMI compared to 20 clinical measurements (r = 0.92)¹⁰. Intermalleolar distance provided excellent inter- and intraobserver variability (r = 0.98 and 0.99, respectively). Along with the 4 other measures of BASMI it takes a total of 7 minutes to complete. Thus this measure passes the OMERACT filter requirement, but is cumbersome to perform in daily practice.

COMPOSITE METROLOGY SCORING SYSTEMS OF SPINAL MOBILITY

BASMI¹⁰. This index was based on a study of 20 measurements performed on 43 patients. From this total metrology exercise, 5 simple clinical measurements were defined that most accurately reflected axial status; they included cervical rotation (goniometer), tragus to wall distance, lateral lumbar flexion, modified Schober's, and intermalleolar distance. The BASMI was tested against the 20 clinical measurements in these patients and an additional 54 patients, with excellent agreement. Interobserver reliability was excellent when 3 physiotherapists examined the same patients (r = 0.96 and 0.99, respectively). It should be noted, however, that the BASMI scores are based on a 0–2 score scale for each of the items, with 0 representing normal mobility, 1 mild to moderate reduction, and 2 severe reduction. A total score of 10 may be achieved if the mobility is restricted severely in all 5 measurements. The BASMI showed sensitivity to change with significant differences noted in treatment groups in the infliximab trial¹⁹. However, the Cohen effect size calculated was only 0.29.

EDASMI⁹. The EDASMI includes 4 measures: cervical rotation (by tape measure), chest expansion, lateral lumbar flexion, and hip internal rotation. The score is based on percentiles, 0 representing higher than 80th percentile, 1 the 60th–80th percentile, 2 the 40th–60th percentile, 3 the 20th–40th percentile, and 4 lower than 20th percentile, allowing a total score of 16. Interobserver and intraobserver reliability were measured in 44 patients by a nurse clinician and a rheumatologist, with excellent agreement (ICC 0.94–0.98 for the total EDASMI and for the individual components). Interestingly, both the EDASMI and BASMI correlated very well with measures of structural damage, and poorly with the Bath AS Disease Activity Index. A standardized response mean of 0.44 was noted using the EDASMI in patients undergoing new therapies.

SUMMARY

A number of spinal mobility measurements have been developed and validated in patients with AS. Several pass the OMERACT filter of truth, discrimination, and feasibility. It is now important to include these measures in therapeutic trials both individually and within indices to determine which combination may be the most useful for (1) routine clinical use, (2) outcome and prognostic studies, and (3) evaluation of new therapies. Moreover, it is important to test these measurements in patients with spondylitis other than ankylosing spondylitis.

REFERENCES

Search PubMed for: