Isokinetic Muscle Testing:
Reflections on Future Venues
Article Outline
Abstract:
This paper presents a brief summary of the background to the development of isokinetic dynamometry, while reflecting upon some future potential research fields.
Among the latter, it is believed that further investigation of methodological issues will continue to be the most central. These would include the design of more advanced testing protocols of both muscle strength and endurance as well as reproducibility of findings derived from their applications. Testing of muscle performance in specific orthopaedic and neurological patient groups will be essential for the enhancement of isokinetics as the standard method in the clinic. To that end, protocols will have to be refined and cutoff values signifying change or stagnation in performance will have to be determined. In addition, issues not directly related to either of these general fields, such as the medicolegal applications of isokinetics for assessing muscular compromise due to pathology or trauma, will be the focus of progressing research.
Key words: isokinetics, dynamometry, methodology, applications
Quantitative evaluation of muscle performance did not achieve widespread recognition and use until the early 70s'. Indeed, muscle strength, the single most representative parameter of the theoretical construct: muscle performance, has been semi-quantitatively (some may argue qualitatively) evaluated using manual muscle testing (MMT) [1]. In the course of the long period since its introduction, this method has been subjected to a thorough and comprehensive exploration. In the light of the findings, its accuracy, sensitivity, specificity, and validity have been sharply criticized. It is for this reason that MMT is considered by authorities in the field to be largely a ‘lost art’[1].
In an attempt to render muscle testing more quantitative and until the advent of isokinetic dynamometry, a number of instruments have been devised that permitted measurement of isometric (static) strength of some muscles. For various reasons, none of these became sufficiently attractive to achieve the status of a standard instrument with the possible exception of the Jamar hand-grip dynamometer [2]. The reason for the latter may lie in the ability of this instrument to fit most adult hands, to demonstrate the relationship between maximal grip strength and the length of the relevant muscles, to validly portray hand muscle functions, which is often of an isometric nature, and to be easily applied [3].
However, an important development in the field of muscle performance measurement took place in the late 60s' with the introduction of the first isokinetic dynamometer (ISD) [4]. Incorporating within a single frame a hydraulic piston, a controllable valve, a lever arm and a load cell, this apparatus was able to offer accommodating but passive resistance to concentric contractions along the angular sector of joint motion. Since the resistance was exerted by a rotating lever arm, muscle strength was for the first time expressed in moment (torque) units (Nm or ft-lb, respectively). Furthermore, the recorded resistance mirrored the maximal muscular moment as soon as the angular velocity of the relevant joint equaled the preset velocity of the dynamometer's lever arm. Clearly, by locking the lever arm, isometric strength could also be measured.
The first generation of the so-called passive ISDs became almost obsolete with the introduction of the active ISD some 15 years later. Its main additional features were an active power source, largely (and currently, exclusively) in the form of an electric motor, and a personal computer. Thus, a modern ISD can measure the isometric, concentric, and eccentric capacity of the muscle under a variety of conditions while performing real-time processing of the force or moment signal in order to present the findings in numerical and graphic forms.
Though no reliable worldwide statistical data is available, it may safely be argued that the number of ISDs located in health and research related facilities may well be counted in thousands. Indeed, so significant has the impact of this technology been on the clinical aspects of muscle performance science that the number of scientific and clinical papers reporting the use of isokinetic dynamometry is quoted in the thousands. Moreover, knee musculature which was initially the main focus of isokinetic dynamometry, has been supplemented by substantial research on the other major muscle groups of the upper and lower extremity joint systems [5] as well as those operating on the trunk.
Even if the only parameter of interest is strength, the number of factors associated with its measurement is staggering. The significance of the aggregate of these factors cannot be overlooked; it constitutes one of the major obstacles to straight-forward application of isokinetic test findings. Specifically, the factors that are involved in every single test relate to the subject (patient), the measurement system, the technique (including the examiner) and the testing protocol. For instance, one could mention age, gender and health status as ‘patient’ factors, the type of the ISD (eg, Biodex, Shirley, NY, USA; Kincom Chattanooga, TN, USA; or Cybex, Ronkonkoma, NY, USA) as a ‘system’ factor, the proximal stabilization method as a ‘technique’ factor, and the reciprocity of the test as a ‘protocol’ factor [5]. However, despite intensive research and widespread effort towards optimization of testing, no accepted guidelines exist, let alone standard procedures.
It should be realized that by its very nature, the production of a given muscular tension level, be it maximal or sub-maximal, is not totally controllable [6]. Similar to other classes of human performance, the ability to voluntarily adjust and, in particular, reproduce an exact level of tension is limited by various neuromotor factors that will reflect the selection of the muscular activation pattern (spatially and temporally), as well as the central and peripheral noise in the system. This noise can bias the transmission and integration of the neural signals at the various junctions such as the spinal cord. Thus, when a subject repeats a given contraction, there is always a measure of inconsistency in terms of differences between consecutive contractions. This bio-error is further magnified by the above mentioned measurement-related factors and hence the likelihood of achieving perfect strength reproducibility for all muscle groups is an elusive objective.
On a different level, the question of the value of isokinetic findings has for many years been a source for criticism, especially among clinicians. It was generally argued that dynamic strength, as measured by various ISDs, bore little relationship to the functional status of the patient following conservative or surgical intervention. Furthermore, using ISDs as muscle rehabilitation platforms was not warranted since functional activities were not isokinetic in nature.
Regarding the first argument, it must be emphasized that function is a multifaceted construct that has strength as one of its components. Thus, although strength may be restituted following injury, components such as reaction time, accuracy, and coordination may not. Moreover, users invariably regard maximal strength as the benchmark despite the fact that most human activities, particularly those performed during normal daily activities, require submaximal strength levels. With respect to athletes, the situation is somewhat different since muscles may indeed be required to develop high moments. However, these high-demand instances most frequently occur during ballistic rapid movements, for which isokinetic measurements may be less suitable. Adding to that the operation of mechanisms like substitution that can partially compensate for strength losses, the correspondence between strength of a particular muscle group and total body function may be small.
As for the second argument, the acquisition of an ISD solely for the purpose of maximal strength conditioning would, be unwarranted in view of the multitude of commercially available appliances specifically suited for this purpose. As mentioned before, a large number of non-isometric activities are ballistic in nature and thus, would not fit an ‘isokinetic profile’. However, due to its accommodating resistance feature, ISDs provide the strongest overloading stimulus along the full range of joint motion and hence would be highly recommended for deconditioned muscles. However, even more important is the fact that modern ISDs include a moment (force)-limiting option which enable clinicians to adjust the minimal and/or maximal muscle contraction levels. Therefore, the use of controlled submaximal exercise is a feature that adds substantially to the value of isokinetic conditioning, but has attracted relatively little attention. Equally important is the possibility within the conditioning process itself, to derive fast, reliable, and accurate answer, to the question of how far the patient is from a given strength value. This would be particularly valuable with respect to muscle–joint systems where the measurement error can be reduced to an acceptable level, eg, the quadriceps–knee complex.
In recent years, research concerning the application of isokinetic dynamometry, has been channeled into three main streams. One relates to methodological issues such as protocol design and reproducibility of test findings. The second refers to applications of this technology to specific subject or patient groups. Other issues that do not fit into these, such as medicolegal applications, make up the third. The methodological stream is undoubtedly central to isokinetics and engulf a large number of questions. It is outside the scope of this article to provide an exhaustive list; however, the following are a few representative issues that may become the focus of advanced research in the coming years.
The error involved in measuring muscular movement under maximal and submaximal conditions is at the core of isokinetics as has recently been highlighted in a paper by Keating and Matyas [7]. As quoted by the authors “Dynamometry measurements provide estimates of strength…if [these are] error-free estimates of subject ability, any change in the a subject's score would indicate a true difference in subject ability in the performance of the test… However, dynamometry does not provide consistent estimates… and, therefore, even when there is every reason to believe that subject strength and test conditions are unchanged, some variability in measurements is always seen on repeated [occasions]”. From an opposite point of view, and more pressing clinically, is the question whether strength differences such as observed following rehabilitation reflect a true improvement (deterioration) in a patient's capacity. It would also be of value to know to what extent bilateral differences, when relevant, are genuine.
A typical isokinetic test of maximal muscular capacity (that follows familiarization and warm-up) consists of three to four consecutive contractions. As mentioned before, within-test variations are discernible even when the test is based on reciprocal (ie, without intercontraction pause) contractions. In such paradigm, a perfect pointby-point overlap among the series of contractions is quite impossible to achieve to the extent that even the peak moment may differ. This non-overlapping (the reciprocal of the error) will not decrease with a longer intercontraction pause, and may actually increase. Since it would be challenging to reproduce exactly the same testing conditions if the subject were to dismount the seat (eg, for knee testing), the error associated with a longer inter-testing period is bound to be even larger.
Confounding this difficulty even further is the effect which learning has on the scores. It would appear that for certain muscle groups and test protocols, learning expressed in improvement not due to actual conditioning, could be quite significant [8]. Therefore, to examine the significance of a change obtained following treatment, it would be imperative to deduct the component attributed to learning. Since this effect on its own is associated with variability, interpretation is even more confounded.
Therefore, the extent of reproducibility reported in terms of correlation coefficients such as Pearson's product moment or intra-class coefficient (ICC) would not suffice for decision making. What is needed is a measure of error that would enable clinicians to decide, based on a cut-off value, whether a change has indeed taken place. Such a measure is the standard error of measurement (SEM) which is expressed in Nm- or ft-lb, and is one standard deviation of the distribution of the error associated with a test score. The SEM is calculated based on a test–retest paradigm. Based on the SEM reported in various sources or calculated from the data presented, Newton et al have indicated that as far as subject related factors were concerned, isokinetic measurements of healthy subjects appeared to be less variable than those derived from impaired subject, and that women had smaller SEMs than men [7]. Test factors such as the movement tested (eg, extension vs flexion) and test velocity also affected the SEM. Most relevant was the observation that variability bore a positive relationship to the magnitude of the muscular moment.
The result of these observations lead one to conclude that in order to provide a clinically meaningful interpretation, the cut-off parameter, be it the SEM or other, must be quoted in terms of the specific population it is based on, and with respect to a given test protocol. With respect to the specific population, future research will have to concentrate on clinical groups with sufficiently common symptoms to warrant the establishment of such parameters (see below). As the magnitude of the mean score affects the SEM, it would stand to reason that when considering a physiotherapeutic treatment process, initial, genuine strength gains may be better identified, whereas those reported later in the process would be less reliable.
Quite on a different front, and relating to the basic methodology of isokinetic-test-finding interpretation, the question of ‘which type of strength’ is the more representative of function has not been properly addressed. This is empirically evident from the nonoverlapping of bilateral strength differences based on either concentric, eccentric, or isometric tests. Research has indeed indicated that dynamic strength differs from static (isometric) strength as shown with respect to forearm [9] and grip [10] muscles. Therefore, when determining a chronic strength insufficiency, the question of which strength type should be used is highly problematic. Currently, no firm guidelines have been offered. Moreover, reliable sources such as the American Guides for the Evaluation of Permanent Impairment [11] elected to ignore modern methods of strength testing in favour of the obsolete method of MMT. That this is the case is alarming enough given the sharp criticism mentioned above [1]. However, this question deserves a thorough investigation and perhaps clinically relevant pooling of the three measures, could better portray the loss. One such example is strength deficiency in the hip abductors versus grip musculature. The previous muscle group operates eccentrically (restraining contralateral descending of the pelvis) and hence determination of its eccentric strength loss due to eg, trauma or neural involvement would be justified. On the other hand, grip muscles contract predominantly isometrically and thus measuring their isometric strength, using an instrument like the Jamar, is clinically and functionally sensible.
Intimately related to this issue is the choice of using derived parameters such as the eccentric to concentric strength ratios. Strictly speaking, the use of strength ratios was erroneous since isokinetic measurements lacked a true zero [12]. This problem has been rectified when ‘gravity compensation’ procedures were introduced in the mid-80s’. Initially, the concentric strength of antagonistic muscles was compared, most commonly with respect to the flexors and extensors of the knee and trunk. However, with the growing use of eccentric testing, the reciprocal mode ratios of the antagonistic muscles was investigated. This reflected accumulated knowledge regarding the restraining effect of one eccentrically contracting muscle group relative to its concentrically contracting antagonist. Thus the dynamic control ratio: Hecc/Qcon where H and Q denote the hamstring and quadriceps groups respectively, has been shown to effectively differentiate between chronic ACL insufficiency and the uninvolved knee muscles in the same patient [13]. Recently, this ratio was studied with respect to the rotatory muscles of the shoulder joint namely ERecc/IRcon, where ER and IR refer to the external and internal rotators respectively [14]. Strong correlations were reported between these ratios and the physiological cross-section area of the supraspinatus. Likewise, deficiencies in the ratio corresponded to malfunction of the rotator cuff in selected pathologies [15].
The correspondence between clinical electromyographic (EMG) and isokinetic measures has not been thoroughly investigated. This aspect of isokinetics may well be a target for future research. Its added benefit would be validation of the system. Clinical EMG is routinely performed in a number of pathologies, however, its ability to render a quantitative assessment of the mechanical dysfunction is rather limited. Thus, when confronted with questions regarding the residual mechanical capacity, clinicians could benefit from conducting isokinetic tests, under varying conditions, of the relevant muscle groups. This void reflects once again the relative ignorance of the medical community with regards to the possibilities afforded by modern ISDs.
Finally, in consideration of the methodological issues, the ongoing research relating to testing of difficult joint–muscle systems is likely to proceed. In particular, reference is made to the hip, trunk, and shoulder musculature for which the range of testing procedures is not only vast but seems to be quite a large distance away from standardization. For instance, testing of trunk musculature is still being pursued in either the seated or upright position resulting in significantly different test findings [5]. Furthermore, since patients impaired with low back dysfunction (LBD) find it more comfortable to be tested while seated, some attachments were designed to incorporate this position. However, testing of normal subjects, particularly for job screening proved probably to be more conveniently conducted while standing. The same applies to shoulder musculature, but in this case the fact the shoulder complex moves as a whole while the arm abducts renders the assumption of joint-axis alignment quite useless. Thus, correction procedures are needed. Another example, the hip muscles, demonstrate the critical need for proper stabilizing frames, the absence of which confounds the results to such an extent that may not be judiciously applied. In this instance, it should be born in mind that the functional position for testing would be that of standing as these muscles operate habitually in gait. However, as the tested limb must be free to move, this means that the weight is born on the other limb, thus introducing balance perturbation. Without proper stabilization, compensatory motion of the whole body is required, length–tension relationships are distorted, and a valid measurement cannot be executed. Surprisingly, some ISDs offer attachments suitable for hip testing in the supine or side-lying position where, particularly for the abductors, the mere attempt to hold the limb horizontally, would be a formidable undertaking for very weak patients. Therefore, the pursuit of better attachments that will eventually result in higher reproducibility will in all likelihood occupy a certain niche in isokinetic research.
As for the second stream, namely, research concerning muscle performance in patients' groups, relatively little has been done. This is regrettable since strength deficiency is found in so many pathological situations, a number of which are chronic (eg, multiple sclerosis and Parkinson's disease) and afflicts a large number of older people. There are a few reasons for this phenomenon including the availability and compliance of patients both for initial and later testing, the time consuming nature of isokinetic testing, the relatively limited number of clinicians skilled in conducting these tests, and the fluctuating nature of strength findings. These all result in large SEMs that may inevitably lead to a decision against the investment of further effort in this most important field. But, nothing could be more wrong with this approach. One refers in particular to the instability of the strength findings [16] that characterize some neurological disturbances in predominantly elderly patients. Since strengthening exercises for afflicted patients may be one of the therapeutic objectives, the ability to establish a reliable baseline as well as the need to follow-up changes are central to the treatment. The fact that strength varies to such an extent among these patients, requires even more research input in this field in order to set rational treatment goals. Likewise, some of the discharge from rehabilitation criteria in young, orthopaedically involved patients, must be based on strength parameters derived from specific pathology-related groups. With the wide spread use of ISDs and multicentre clinical trials, there is no reason why in the near future, effort on a national and international basis is not undertaken in this respect.
Within the third stream, one could identify research in isokinetics which does not necessarily fall within the confines of the above two streams. One particular example is the use of ISDs in identifying submaximal performance, an area that has immense implications in medico-legal studies of muscle weakness.
Muscle weakness due to traumatic injury is well recognized by most Western legal system as an indemnifiable damage. As an integral part of the litigation process, patients presenting with injury to the muscular apparatus are requested to be tested in order to assess the extent of damage. Since muscle testing requires collaboration on the part of the patient and in view of the poor validity of MMT [1], especially when strength loss is partial, the question as to whether the patient is indeed performing maximally is often raised by the examiner. The question of maximality or sincerity of effort is, therefore, crucial to the decision since the sums paid to claimants are very large. For instance, in the United States alone, some 70 billion dollars were paid for musculoskeletal injures at the beginning of the 90s'. Hence, even if muscular related injuries constitute only a small fraction, the total amount would still be enormous.
The basis for quantifying the loss was compared with the same muscle group in the uninvolved side (extremity). This entailed instrumental measurement of strength that was invariably isometric. In order to test the sincerity of the effort, the prevailing approach, until recently, was linked to the principle of consistency. According to this principle, only maximal contractions that correspond to sincere effort may be repeated with relatively small inter-contraction variation (namely high consistency). To quantifying the consistency, the coefficient of variation (CV), which is a percent related index expressed by the following formula: CV = (standard deviation of the scores/mean value of the scores) × 100 was applied. Alternatively, submaximal (insincere) effort would result in relatively large CVs. Small and large values were arbitrarily assigned; for instance it was suggested that a CV of up to 10% indicated a sincere effort whereas anything above was considered as indicative of feigned effort [17]. This, or near cut-off values, were also adopted in isokinetic measurements of strength. However, an increasing number of studies conducted in recent years [17] that compared the CV obtained during maximal performance with that which was derived when subjects were told to feign weakness, have decisively indicated that the CVs was not an effective index. Indeed the sincere CV was significantly higher than the nonsincere one, but there was too much of an overlap between the two. This in turn, resulted in an unacceptable rate of false positives and/or false negatives which was reflected in relatively low specificity and sensitivity. Thus, the use of the CV in sincerity of effort testing was effectively rejected.
Based on application of isokinetic dynamometry, a different approach has been adopted in a series of studies relating to the identification of submaximal effort in normal subjects using various muscles [18, 19, 20, 21]. Instead of looking at the consistency of performance, the compatibility of the test findings with the expected variations in the force–velocity (F–V) relationships of skeletal muscle was sought. It was argued that since the concentric branch of the F–V curve decreased hyperbolically with increasing test speed while the eccentric branch remained more or less constant, the eccentric/concentric (Ecc/Con) strength ratio (either peak or average torquebased) should increase with the speed. Specifically, if performed under low and high testing speed, the following index: DEC = (Ecc/Con)high–(Ecc/Con)low (where high and low refer to the test speed) would behave differently in maximal versus submaximal efforts. This assumption was based on the fact that eccentric contractions being defensive in nature were less amenable to feigning, increasingly so with an imposed rise in the test speed.
As the above studies have strongly indicated, the DEC proved to be highly effective in differentiating the performance level. This was expressed by a conspicuous separation between the frequency distributions of the DECs corresponding to sincere and insincere efforts. It should also be emphasized that in these studies as well, the CVs relating to the sincere and insincere performance behaved in much the same way as reported in previous research. Furthermore, in a recent study, the test protocol was employed on two occasions and in spite of a discernible learning effect in 95% of the subjects, feigning of weakness could still be identified in the second test session [22]. Since this study was based on the use of extremely short range of motion, it has also opened a new window into utilization of different isokinetic measurement systems, namely the use of linear motion rather than angular motion based systems. Indeed, in years to come, new designs may be introduced that will pave the way for the formulation of new concepts of muscle strength production, controls and assessment.
References
- . Muscle performance evaluation in orthopaedic practice. J Bone Joint Surg Am. 1990;72:1562–1574
- . Grip test: the use of a dynamometer with adjustable handle spacings. J Bone Joint Surg Am. 1954;36:820–824
- . A study of isokinetic exercise. Phys Ther. 1969;49:735–741
- . Isokinetics: Muscle Testing, Interpretation and Clinical Applications. Edinburgh: Churchill Livingstone; 1995;
- . Motor Control and Learning: A Behavioral Emphasis. Third ed.. Champaign, IL: Human Kinetics; 1999;
- . Unpredictable error in dynamometry measurements: a quantitative analysis of the literature. Isokinetics and Exercise Science. 1998;7:107–122
- Trunk strength testing with iso-machines part II: experimental evaluation of the Cybex II back testing system in normal subjects and patients with chronic low back pain. Spine. 1993;18:812–824
- Isometric and isokinetic endurance testing of the forearm complex. Am J Sports Med. 1991;19:107–111
- . The measurement of isokinetic fingers flexion strength. Clinical Biomechanics. 1998;12:473–481
- . Guides to the Evaluation of Permanent Impairment. 4th ed.. Chicago: American Medical Association; 1993;
- . Clinical uses of isokinetic measurements: critical issues. Phys Ther. 1987;67:1840–1844
- Thigh muscle activity and anterior cruciate ligament insufficiency. Clin Biomech. 1989;5:68–72
- David G, Magarey M, Jones MA, et al: EMG and strength correlates of selected shoulder muscles during rotations of the glenohumeral joint. Clin Biomech (Submitted 2000, in press).
- . Shoulder strength and range of motion in symptomatic and pain free elite swimmers. Am J Sports Med. 1997;25:454–460
- Reproducibility of isokinetic muscle strength measurements in normal and arthritic individuals. Scand J Rehabil Med. 1990;22:93–99
- . Review of sincerity of effort testing. Safety Science. 1997;25:237–245
- . Suboptimal muscular performance: measuring isokinetic strength with a new test protocol. Arch Phys Med Rehabil. 1996;77:578–581
- . Differentiation of submaximal from maximal trunk extension effort. Spine. 1997;22:2672–2676
- . An isokinetic study of submaximal effort in elbow flexion. Percept Mot Skills. 1997;84:1431–1438
- . Identification of submaximal grip effort using isokinetic dynamometry. Clin Biomech. 1999;14:522–527
- Dvir Z, Keating JJ. Identification of submaximal trunk extension effort using extremely short range of test motion. (Submitted 2000).
PII: S1013-7025(00)18002-9
doi:10.1016/S1013-7025(00)18002-9
© 2011 Elsevier B.V. All rights reserved.
