Range of Motion (RoM) Measurements

Today’s method for conducting functional testing of the spine (flexion/extension and lateral bending radiographs, which are referred to in this text as the clinical standard of care) involves capturing standard radiographs of the spine as subjects bend, and then hold their spines fixed in the extremes of motion in either the sagittal (in the case of flexion/extension) or coronal (in the case of lateral bending) planes. These studies are separate to, but often used as an adjunct with, other medical imaging studies such as plain radiographs or CT scans in the diagnostic assessment of a patient’s spine. When performing these motions, each subject bends in each direction to his or her own maximum voluntary bending angle (MVBA).

These two images taken at the extremes of trunk bending within a single plane are then interpreted — either manually using a pen, ruler, and protractor or more recently, with the advent of digital imaging, an imaging workstation — to derive range of motion (RoM) measurements. RoM measurements represent the total displacement between any two vertebrae during MVBA bending, and are expressed as both angulations, as measured in degrees and referred to in this text as the intervertebral angle (IVA) in either the coronal or sagittal plane, and translations in the sagittal plane, measured in millimeters and referred to in this text as the intervertebral translation (IVT). See Figure 10-1 for a simplified diagram showing how IVA and IVT are derived from radiographic images.

FIGURE 10-1 Simplified diagram of how IVA and IVT are derived from radiographic images.

RoM is defined by the rotation of the body (IVA) and the translation of a point on the body (IVT). While the rotation is unambiguous, the translation is not. The translation is different for different points of the vertebral body and, additionally, it is subject to magnification and distortion on radiographs. This ambiguity has led to: (1) the introduction of multiple techniques for selecting points on the vertebral body and measuring IVT;^2,4,8,21,22 (2) attempts to define standardized displacement thresholds for what constitutes translational instability;⁹ and (3) the proposal of multiple systems for scoring and classifying translational instabilities (there have been the Myerding scale,¹⁰ the Newman Scale,¹¹ and the modified Newman scale¹² for scoring translational instabilities, as well as the Wiltse¹³ system for classifying them).

Despite the multiplicity of different methods that have been proposed over the years, the Myerding system has become the most widely used in clinical practice and has thus emerged as the standard system by which translational instability is graded. The Myerding system categorizes the severity of a translational instability based upon IVT measurements expressed as a percentage of the total superior vertebral body length (also measured in millimeters): grade 1 is 0% to 25%, grade 2 is 25% to 50%, and grade 3 is 50% to 75%;

Grade 4 is 75% to 100%; over 100% is spondyloptosis, when the vertebra completely falls off the supporting vertebra. One key advantage of the Myerding system is that it is a relative grading system, meaning that it helps to control for distortion and magnification errors that can be associated with absolute measurements of displacement (millimeters) derived from radiographic images.

Although IVT measurements have been the subject of intense investigation over the years, it is not a topic about which there is currently much debate. This topic was thoroughly explored in studies published in the 1970s through 1990s; however, in the past 15 to 20 years a de facto consensus has emerged with respect to the use of the Myerding system as the clinical gold standard for grading translational instability cases. The same is not true for IVA measurements, as no consensus has emerged with respect to the clinical application of IVA despite a very large volume of recent investigational activity. Therefore the remainder of this chapter will present a review of past and current knowledge with respect to IVA, with a particular focus on patient-management implications for treatment of the aging spine.

IVA is used clinically to assess intervertebral articulation in either the sagittal or coronal planes, and as such should theoretically be capable of detecting six specific types of intervertebral functional presentations (see Figure 10-2):

FIGURE 10-2 Theoretical framework for the detection of six functional presentations based on IVA measurements.

However, there is a large gap between those six presentations that should theoretically be detectable, and those that are actually detectable with the current clinical standard of care. This gap is thoroughly explored in the following sections, and must be understood by the clinical practitioner in order to properly interpret functional testing results.

Measurement Variability in Range of Motion (RoM) Measurements

As with any quantitative diagnostic measurement parameter, measurement variability is the key driver of diagnostic efficacy in the application of such measurements to differentiate between the various types of patient presentations. Simply stated, measurement variability is the enemy of effective diagnosis: the higher the measurement variability, the less effective the resulting diagnosis. In the case of RoM measurements, it has been shown that measurement variability is high^2,3 and diagnostic efficacy is low.^4–7 The causes and effects of this measurement variability are well understood; however, the implications for the clinical practitioner have rarely been discussed in the published literature. Therefore one of the main goals of this section is to present a data-driven analysis of RoM measurement variability and how this variability should be taken into account in the interpretation of functional testing results used in the diagnosis of spine disease and management of the aging spine.

RoM measurement variability is composed of variability between/within observers, and variability between/within subjects. Variability between observers is referred to as interobserver variability, while variability associated with a single observer taking multiple measurements at different points in time is called intraobserver variability (also called test/re-test variability). Similarly, variability between patients is referred to as intersubject variability, while the variability of any given patient between multiple tests taken at different points in time is referred to intrasubject variability. For example, intersubject variability can include the effects of physiologic differences from patient to patient, whereas intrasubject variability can include variability in the willingness of a patient to perform bending motions from test to test (which can often be due to the influence of pain and/or fear of pain among other things).

There is also a third component of RoM measurement variability that relates to the variability that exists between different testing sites. Different testing sites utilize different radiography platforms, and different imaging platforms can produce different types of image distortion, magnification, and other image variants. Further, different sites utilize different practices for patient positioning and image analysis. These variations among testing sites can directly contribute to RoM measurement variability and therefore must also be taken into account. For the purpose of this discussion, this variability among different testing sites will be referred to as intersite variability.

The different types of RoM measurement variability mentioned in the preceding paragraphs are interrelated in several ways that can be best understood through the concept of “accumulating” variability. As previously discussed, intra-subject variability is a measurement of the test/re-test variation within a given subject, while inter-subject variability is a measurement of the variability across a population of subjects. However, since the RoM measurement from any given subject is affected by intra-subject variation, then any measurement of inter-subject RoM variability across multiple subjects would necessarily “accumulate” the combined effects of intra-subject variation and inter-subject variation. The same concept holds true for measurements of inter-observer RoM variability, namely that these measurements accumulate the effects of both intra-observer and inter-observer variation.

This concept of “accumulation” of variability also applies to the overall relationship between observer-related variability (interobserver and intraobserver variability) and subject-related variability (intersubject and intrasubject variability). Subject-related variation in intervertebral motion exists as an inherent property of the physiology of the spine. In other words, there is a certain amount of variation that is inherent to the way the spines of different people move, or in the way a given person’s spine moves at different points in time. For this discussion, we will refer to this inherent variation as the “pure” intrasubject and intersubject variability. However it is impossible to measure this “pure” intrasubject and intersubject variability without constructing an observational system to take measurements, and any observational system constructed to take measurements is also subject to both intraobserver and interobserver variability. Therefore any measurement of intersubject variability, for this discussion called “observed intersubject variability,” necessarily “accumulates” the combined effects of both observer-related variability and subject-related variability.

See Figure 10-3 for a simplified conceptual diagram of how selected types of RoM measurement variability interrelate through the accumulation of measurement variability.

FIGURE 10-3 Simplified conceptual diagram of the “accumulation” of RoM measurement variability, which applies to both IVA and IVT measurements. Note that this diagram is considered simplified because it does not represent every possible type of measurement variability. For example, observed intrasubject variability is not represented. This simplified diagram represents the interrelationships between those types of measurement variability that are most important for the clinical practitioner to understand in evaluating the performance of today’s in vivo methods of spinal functional testing.

Using Normative IVA Data to Detect Normal Motion, Hypomobility, and Hypermobility

As previously discussed, it is theoretically possible to use normative IVA data from a population of asymptomatic subjects to differentiate normal from hypomobile and hypermobile intervertebral motion (see Figure 10-2). However, with the current standard of care for conducting spinal functional testing, only hypermobility and pseudarthrosis can be detected with an acceptable level of statistical confidence. This fact, although not widely discussed, has very significant implications in terms of patient management, which are discussed later in this section. However as a starting point to this discussion, it is necessary to first re-examine the conventional wisdom regarding what is currently considered  “normal healthy” intervertebral rotation.

As a general biostatistical principle, a quantitative diagnostic value is considered an outlier and therefore abnormal if it lies above or below two standard deviations of the mean value that is observed among a representative sample of normal healthy subjects (the mean plus and minus two standard deviations represents approximately 95.5% of all observed values). Therefore, the magnitude of such standard deviations will determine the specific ranges or IVA that should be considered normal versus hypomobile or hypermobile. Many investigators over the years have conducted studies of IVA values across asymptomatic populations for the purpose of producing such ranges, yet all of these investigators are plagued by the same Achilles’ heel: they are all single-site studies and therefore fail to account for intersite variability. Thus every single-site study underestimates IVA measurement variability and therefore produces unreliable ranges of what constitutes normal versus hypomobile or hypermobile intervertebral rotation. However, by conducting a meta-analysis of these studies it is possible to account for this intersite variability and produce more representative ranges of what constitutes normal IVA.

In conducting this meta-analysis, a total of 22 published IVA datasets were identified (15 lumbar and 7 cervical). Each dataset was carefully examined and screened to ensure that: (1) the method for measuring IVA was consistent with the current clinical standard of care, and (2) the variability (standard deviation, or SD) among observed IVA values was published along with the mean. After applying this screen, three lumbar datasets and four cervical datasets qualified for this meta-analysis. See Table 10-1 for a list of all 22 datasets that were considered.

After including all qualifying datasets, the following values were tabulated for the mean and standard deviation of observed IVA values taken from multiple populations of asymptomatic subjects across multiple sites (Table 10-2). The standard deviation values in the “Aggregated Across Sites” column at the far right of each table represent the standard deviation of the superset created by combining the observed values from all sites, and represents the observed intersite variability associated with the current standard of care for measuring IVA at each level, while the standard deviation values for each investigator represent that investigator’s site’s observed intersubject/intrasite variability.

Using these normative values that account for the effects of intersite variability, it is possible to produce threshold IVA values that represent hypomobility and hypermobility, as given in Table 10-3.

Effects of IVA Measurement Variability on the Diagnostic Efficacy of Functional Testing of the Spine