GAIT EVALUATION AND MANAGEMENT: INTRODUCTION
The objective of this chapter is to present an overview of clinical gait analysis (GA). GA is a process of instrumental measurement and evaluation of walking ability in patients with specific problems related to locomotion. This analysis is intended to provide answers to specific clinical questions, which in turn affects future clinical decisions and monitoring of the patient. GA is often conducted in a laboratory setting by clinicians with subspecialty training; its goal is also to aid in functional restoration.
Cerebral palsy represents an example of a disease state in which clinical GA has provided advanced management strategies that substantially affected functional restoration. An accreditation in GA helps to ensure that quality standards are met. Most common measures taken during GA include time–distance parameters; kinematics of body segments and joints; kinetics, an analysis of the forces that come into play during motion; muscular performance; and energy consumption. Generally, data obtained are summarized in a clinical report, including their biomechanical and clinical interpretation and treatment recommendations.1
Gait is the most important method of human locomotion characterized by periods of loading and unloading of the limbs to facilitate movement and function. Such movement affects many other functional activities such as performance of activities of daily living (ADLs), involvement in social activities, and his participation in a vocation/occupation.
GA consists of an analysis of joint kinematics, kinetics, and dynamic electromyography (EMG) data. The analysis is performed and properly interpreted by experienced individuals and improves basic knowledge regarding an individual’s physiological and pathological gait function.
An accurate understanding of appropriate terminology in the analysis of gait movements is of great importance. Different biomechanical and muscular variables are often studied and many of them are highly interrelated.2 Clarity regarding the method by which measures are acquired and calculated, units of measurement used and the normalization of standards is critical. Only an accurate and systematic approach to “data handling” will permit the assessment of the efficiency of the instruments and methodologies of acquisition and evaluation.3
THE GAIT CYCLE
The “gait cycle” is defined as the set of moments and events that take place between two successive initial contacts of the ground with the same foot. It can be divided into two principal phases: the “stance phase,” during which the foot remains in contact with the ground, and the “swing phase,” during which the foot is brought forward. In normal subjects, the cycle begins with a heel-strike (0%), and terminates with another heel-strike of the same foot (100%). When a person is walking at normal speed, the stance phase typically constitutes about 61% of the entire cycle, and the swing phase takes up 39%; the various events in the gait cycle are normally expressed as a percentage of the whole cycle, which normalizes them to facilitate the comparison between subjects. The duration of a cycle naturally varies with velocity and is around 1.1 s.
The alternation of the stance and the swing phases of the two limbs create a further subdivision of the cycle into single-limb support and double-limb support phases. The double-limb support phase is repeated twice during the cycle, at approximately 10% and 50% of the cycle, and its duration is about 20% of the cycle. In normal subjects, the average duration of the walking cycle and the duration of the stance and swing phases must be equal on both sides, while in pathological gait, the values of the sides may differ, giving rise to arrhythmic gait.
Traditionally, the gait cycle has been described with respect to the movement of the foot and corresponds to the gait of normal subjects. However, there are a number of patients with pathologies (e.g., ankle equinus secondary to upper motor neuron disease), whose gait cannot be described using this approach.
In the traditional nomenclature, the stance phase events are as follows (Fig. 97–1):
Heel-strike or heel contact, initial contact, foot strike: Initiates the gait cycle and represents the point at which the body’s center of gravity is at its lowest position
Foot-flat: Is the time when the plantar surface of the foot touches the ground
Midstance: Occurs when the swinging (contralateral) foot passes the stance foot and the body’s center of gravity is at its highest position
Heel-rise or heel off: Occurs as the heel loses contact with the ground and pushoff is initiated via the triceps surae muscles, which plantar-flex the ankle
Toe off: Terminates the stance phase as the foot leaves the ground
The swing phase events are:
Acceleration: Begins as soon as the foot leaves the ground and the subject activates the hip flexor muscles to swing the leg forward
Midswing: Occurs when the foot passes directly beneath the body, coincidental with midstance for the other foot
Deceleration: Describes the action of the muscles as they slow the leg and stabilize the foot in preparation for the next heel-strike.
An alternative nomenclature was developed by Perry and her associates at Rancho Los Amigos Hospital in California.5 There are eight events, described in terms of functional objective to fulfill during gait (Fig. 97–2):
Initial contact: The principal objective of the body in this phase is to position the foot correctly as it comes into contact with the floor (0% of the cycle). The body is about to begin deceleration.
Loading response: From heel-strike to toe off of the contralateral limb (0%–10%); this corresponds to the first period of double-limb support. It is a phase of deceleration during which the shock of the impact of the lower limb on the ground is absorbed.
Midstance: From toe off of the contralateral foot to heel-rise (10%–30% of the cycle); during this phase, the body center of mass (COM) reaches its zenith over the base of support and forward velocity is at its minimum.
Terminal stance: From heel-rise to heel-strike of the opposite side (30%–50% of the cycle); during this phase, the body’s COM moves in front of the base of support and accelerates in such a way that the body falls forward toward the unsupported side.
Preswing: From heel-strike on the contralateral limb to toe off on the stance side (50%–60% of the cycle). The principal goal of this phase is to prepare the limb for swing.
In the swing phase, the swinging limb behaves as a compound pendulum, and therefore the duration of swing depends on the mass moment of inertia of its segments. The swing phase can be divided into three periods: a period in which the velocity may be altered (accelerated or decelerated), a transition period, and a final period in which swing velocity may be reversed.
Initial swing: This begins with toe off (60%–70% of the cycle). The principal purpose of this stage is to establish a suitable distance between the foot and the ground and permit a variable cadence.
Midswing: From maximum knee flexion to contralateral heel-rise (70%–85% of the cycle). The principal purpose of this stage is to maintain a suitable distance between foot and the ground.
Terminal swing: From contralateral heel-rise to heel-strike (85%–100% of the cycle). The purpose of this stage is the deceleration of the leg and correct prepositioning of the foot for contact.
The principal goal of locomotion is to propel the body forward. Typically, developed individuals produce a repeatable gait pattern that is both cyclical and symmetric. The description of time and distance variables is crucial in GA, as they are of recognized clinical relevance in the assessment of motor pathologies.
The temporal-spatial parameters are mostly used for screening (e.g., detecting elderly people at risk of falling), performance measuring (e.g., grading patient’s level of disability), monitoring (i.e., as an outcome measure), and normalizing other gait measurements (to compare results from people walking at different speeds).6
The distance between two consecutive supports of the same part of the foot for each of the two limbs is called “step” (Fig. 97–3). “Step length” is the distance on the plane of progression between the support of one part of the foot (in normal conditions, the heel) and the support of the same part of the contralateral limb. The term “step dimension” refers to the limb that is moving forward. In normal walking in a straight line without obstacles, the length of two steps should be the same, and when one of each has occurred, the person has taken a stride, or performed one gait cycle, and the time it takes for this to occur is called the “gait cycle duration” or “stride time.” “Stride length” is the distance between two successive heel-strikes of the same foot; it is equal to the length of the foot plus the distance covered during the swing. “Step width” is defined as the medial-lateral distance between the heels in double support and tends to increase with disequilibrium in order to increase the base of support.
The number of steps per minute is called “cadence,” which is related to the length of the lower limb, in a similar fashion to a pendulum: longer legs have slower cadence. Consequently, most people seem to maintain a constant “walk ratio” (stride length divided by cadence) throughout their lives. Because women are, on average, a little shorter than men, they tend to have a slightly higher cadence.10 Small children have an even more rapid cadence, which gradually falls as they grow taller. The preferred “walking speed” is the speed at which humans or other animals choose to walk. In the absence of significant external factors, humans tend to walk at about 1.4 m/s. Average velocity can be expressed as a percentage of height in order to compare results from people walking at different speeds. Although everyone has a preferred walking speed, the actual speed is continuously adjusted according to the conditions. A person’s natural gait is very dependent on the environment; for example, people tend to walk faster on a long walkway and slower in a short one, and are also influenced by the size of the room.11–14 Consequently, outdoor studies invariably report higher speeds and stride lengths than indoor studies.14–17
Walking speed is related to both cadence and the stride length, so it can be increased by a more rapid cadence, longer stride length, or both. In healthy people, both parameters increase with speed. Cadence increases linearly, and stride length increases logarithmically, changing at low speeds, but tends to level off at higher speeds.
The average onset of walking in children is at about 11 months of age, and walking appears to stabilize at around 4 to 5 years of age.18,19 Natural walking speed remains relatively stable until the age of 70; it declines about 15% per decade thereafter.20,21 Healthy subjects can increase their speed by as much as 44% above the natural, self-selected pace.22 Maximal speed declines earlier and more steeply with age—about 20% per decade after the age of 50. Cadence does not change with age (maintaining its relationship with limb length), so stride length must be the source of the decreased speed.23 As people age, balance slowly deteriorates, and this is reflected in the temporal-spatial parameters, showing reduced stride length and reduced speed; this seems to be associated with fear of falling rather than falling itself.13,24
Sometimes the assessment of step length differences through the analysis of the step length ratio (SLR) is a useful measure of symmetry for tracking a patient’s progress after treatment, with the ratio rising closer to 1 as the gait improves. Other symmetry indices have been described, using variables such as right and left stride times; such measurements allow patients a simple overall outcome measure that can inform the patient of her or his progress during gait training.25–27
Quantitative analysis of kinematics involves the acquisition and numerical elaboration of the variables that define the movement in space of the various body segments, without taking into consideration the forces that determine this movement. These variables are the trajectories of the single segments, the angular displacement, velocities, and accelerations.4 The system of spatial reference can be absolute or relative, and in the latter case, the anatomic point of reference must be specified. It is thus possible to describe the angular rotation of the main joints of the lower limb as motion between segments (relative variables) and the segments’ displacement (i.e., motion of the pelvis or trunk) in space (absolute variables).6 For this purpose, stereophotogrammetric systems, based on infrared cameras, are utilized. From a biomechanical point of view, in order to model the human body to calculate its motion, it is necessary to refer to mathematical rules.
Body segments are considered as a series of rigid segments moving each other in the three planes of the spaces according to general anatomical axes. Markers are positioned on bony landmarks of each body segment to define a plane representing the segment itself for mobilization. Generally, in GA practices, segment anatomical systems are described for the pelvis, femur, tibio/fibula, and foot segment, and in some cases, the trunk, upper limbs, and head are included as well. The joint system for reconstruction of the position and orientation of the lower limb bones in space during gait that is widely used and accepted, according to current international standards (International Society of Biomechanics), is the Joint Coordinate System (JCS).28
The biomechanical model and the procedures for gait data collection (marker set; that is, the number of markers and sites on which they are placed on the segment), processing analysis and reporting of results are generally structured in protocols that allow for making kinematic and kinetic measurements clinically comprehensible.29–32 Several protocols are currently available, which differ considerably for the marker-sets, the variables measured, the degree of freedom assigned to joints, anatomical and technical references, joint rotation convention, and terminology for the biomechanical model implemented.
The Newington model is the first protocol published and the most commonly used technique for gait data acquisition and reduction.33,34 It has been also the basis of many commercial software packages, the most recent being Plug-in Gait (PiG—Vicon Motion Systems, Oxford, UK). The distinction between internal anatomical landmarks and external technical markers was proposed in the 1980s.2 Based on this Calibration Anatomical System Technique (CAST) approach, a number of definitions for reference landmarks and frames, together with overall standard protocols, have been proposed, such as Total3Dgait protocol (or IORGait, C-Motion, Germantown,MD, USA) (see Fig. 97–4).35–37
Frontal and back diagrams of the marker set according to the IOR Gait protocol (Leardini 2007): dark circles indicates the markers to be removed after the static calibration, which are necessary only for the anatomical landmark calibration. On the trunk and column, the marker set according to the relevant specific protocol (Image used with permission of C-Motion, Germantown, MD).
Another issue to consider when performing GA is the precision and accuracy of measurements.38 These are influenced not only by the instrumentation used, but also by the interposition of soft tissues between markers and bones, which have unpredictable effects.39–41 In addition, there is natural intrasubject variability particularly associated to different walking speeds, as well as age, gender, and body mass index (BMI).42–47 Finally, intraexaminer and interexaminer gait data variability may exist. Such variability results from inconsistent bony landmark identification and marker positioning.48,49
The main measurement gathered for lower limb joints are usually pelvic rotations (tilt, rotation, and anteversion/retroversion) and the movement of the hip (flexion/extension, adduction/abduction internal/external rotation), the knee (flexion/extension, internal/external rotation), and the ankle-foot complex (dorsiflexion/plantarflexion). Because most protocols do not calculate the motion of the foot on the three planes of space, the foot angle was introduced in order to take a measurement of the position of the foot with respect to the line of body progression. More recent protocols have introduced the full three-dimensional (3D) motion of the ankle-foot complex, including the inversion/eversion and the abduction/adduction angles (Fig. 97–5).50 Furthermore, dedicated protocols for the foot kinematics based on a number of foot segments are available.51
Example of 3D ankle-foot complex kinematics in a patient operated on for equinus-varus foot after suffering a stroke. Upper row: Sagittal, transverse, and coronal kinematics: gray lines right foot (healthy side), black lines (paretic side), and gray band: normalcy. The lack of a flexion-extension pattern and the inverted-adducted attitude is evident. Lower row: After surgical intervention for equinus-varus correction, the kinematic pattern is well restored.
The term kinetics describes the measurement and the analysis of the forces, the power, and the energy developed during movement.1 Evaluating the role played by the various forces (i.e., gravity, muscle, and inertia) in the various phases of the gait cycle is very difficult. One method, which has been adopted in the majority of gait laboratories, consists of the measurement of the reaction force of the ground to the force exerted by the stance foot (the ground reaction force, or GRF).
A comprehension of GRF is important in GA. Analysis of GRF helps clinicians to understand the relationships of joint movements (e.g., anterior, posterior, superior, or medial) to the center of rotation. A critical concept in GA is to understand that the GRF is related to mass, with the smaller mass of the two masses in contact having a larger acceleration. GRF is often characterized by vector with magnitude and direction and is measured by force plates in a GA lab. For this purpose, a dynamometric platform is typically utilized.4,52 This platform records the forces exerted by the subject walking on it and transforms them into an electrical signal that can be displayed in real time in the form of vectors corresponding to the GRFs in the three spatial planes.
The succession of these vectors during stance at fixed intervals determines a figure defined as the “vectorial-butterfly diagram,” which in normal walking has typical characteristics (Fig. 97–6).53,54 Such diagrams of the two feet of the same subject should be more or less the same, and these are repeatable at the same speed. The length of the vector represents the quantity of applied force, expressed in Newtons, or normalized in percentage of body weight; the direction of the vectors is equal to and opposite of the direction of the applied forces.7
The trend of moments and powers in normal subjects is one of the parameters that are most susceptible to variation according to the methods of acquisition and calculation that are adopted.
In fact, joint moments are usually calculated through the relationships of vertical GRFs and lower limb joint center of rotation.7,8 Inertial parameters also can be considered.7 Joint moments can be presented as internal (the forces that muscles and passive soft tissues, ligaments, and capsules exert) or external (representing the GRF’s action on joints).6 Joint moments have been shown to provide very useful indirect information, as they demonstrate muscle activity in relation to joint motion during gait. Furthermore, to determine the purpose of muscle contraction, another approach is necessary: power analysis. The flow of power through the limb provides insight into the source and destination of the power responsible for driving the gait pattern.
During gait, muscle is capable of three basic functions: shortening against a load (concentric contraction), lengthening against a load (eccentric contraction), and maintaining constant length against a load (isometric contraction).6,55 During a concentric contraction, the muscle generates power, while in an eccentric contraction, it absorbs power; meanwhile, if the power is absent, no muscle can be contracted. Agonist and antagonist muscles contract in order to stabilize the joint.56,57 From a clinical point of view, power is probably the single most informative biomechanical variable.57
MUSCLE ACTIVITY (DYNAMIC EMG)
The purpose of dynamic EMG in clinical GA is essentially to define the muscular activity that controls joint movement during gait, as shown by numerous studies carried out on children with cerebral palsy (CP), Abnormal patterns of muscle activation are used as an indication for surgical tendon transfer or lengthening.58,59
The correlation between muscular events and the study of moments and joint power may provide further data of clinical interest on the role of muscle control during gait in relation to biomechanical and rehabilitation problems.60,61 The characterization of muscular activity, however, is a very delicate process because many factors come into play, and they make clinical interpretation difficult.62 The electromyography (EMG) signal is an electric signal associated with the contraction of a muscle and the signal is produced by the depolarizing of motor units, often referred as “motor unit action potential (MUAP).”
Muscle force is generated by the summation of the tiny forces generated in each muscle fiber. Increase in force arises from recruitment of more motor units (i.e., spatial summation) and increased firing rate (i.e., temporal summation) of motor units. The summation of many MUAPs from all the motor units active at a given time results in myoelectric activity called the “electromyogram.” This can be picked up over the skin surface over the muscle (surface EMG) or by percutaneous (indwelling) fine needle electrodes inserted into the muscle belly.63 Recording EMG activity using surface electrodes requires care in electrode application and attention to the recording details. The exact take-up area depends greatly on characteristics of the surface electrodes (e.g., diameter, interelectrode distance) and the signal quality is conditioned by the characteristics of the skin (e.g., the presence of hair and sweat).
Guidelines for appropriate electrode positioning and skin preparation have been published.64 Numerous visual techniques are available for presenting EMG data during gait. Frequently, simple presentations of the raw data may be an important way to provide information to the reader. The envelope of the myoelectric signal historically has been used to estimate the intensity of activation and the measurement of muscle intervals of activation during the gait cycle. With regard to the extraction of the envelope, the techniques for integrating, rectifying, and averaging the signal are numerous, and there are no universally accepted criteria for the solution of technical problems (e.g., time constants, filter characteristics, or number of repetitions for averaging).65
Also, the study of envelopes in a clinical context should be normalized to the maximum voluntary contraction (MVC) or scaled to the maximal walking signal.66 Such analysis provides an estimate of the muscle force exerted during the dynamic contraction. However, the amplitude of the myoelectric signal recorded during dynamic contractions depends on several physiologic, anatomic, and technical conditions. Therefore, the correlation between the instant value of the envelope and the force exerted is questionable.62,66 Information about the shape of the envelope, its amplitude, the position and sharpness of the peaks of myoelectric activity, and its frequency content is nevertheless useful for studying muscular function under normal and pathological conditions.
The study of intervals of muscular activation has been widely used for clinical purposes. Using the computed onset and offset durations, periods during which each lower limb muscle is active can be presented along the gait cycle and be interpreted in terms of the function that each muscle exerts in different gait phases.
As one might expect, the EMG activity from each muscle can depend on walking speed; consequently, it is important that walking speed be controlled for each subject. Also, there are considerable intersubject differences in the normal gait profile; therefore, it can be difficult to establish a normal gait pattern. More recently, it has been recognized that there may be important variations from stride to stride within an individual subject. Recently, the use of statistical analysis of EMG data about a high number of consecutive strides has opened new possibilities for interpreting normal and pathological variability (Fig. 97–7).67
Example of statistical analysis of the EMG signal (used with permission of M. Knaflitz and V. Agostini). In the upper graph, the raw EMG signal of three typical activations of the tibialis anterior is presented. The lower graph represents the statistical analysis of the frequency (mean and SD) of activation of the tibialis anterior muscle during several consecutive strides.
THE DETERMINANTS OF GAIT AND LOCOMOTOR FUNCTIONS
One of the pioneers of GA, Verne Inman, thought that the driving aim of gait was to minimize vertical and horizontal motion of the COM to maximize efficiency (Fig. 97–8).3 Potential energy would have to be used to raise the body each cycle, which would be wasteful. Inman theorized that COM motion would be reduced by increasing the effective lower-limb length during double support and decreasing it during single support. Six mechanisms or determinants were described.68
Pelvic rotation: The pelvis rotates forward (internally) on the leading leg to increase its effective length. It also rotates externally on the contralateral side to increase the length of the trailing limb simultaneously.
Pelvic obliquity: The pelvis lists downward to increase the effective leg length of the trailing limb. There is a problem with this determinant, as subsequent analysis of the pelvis kinematics demonstrates that the timing is not correct—the pelvis does list downward, but too late to help raise the COM.
Stance phase knee flexion: The knee flexes during single support to effectively shorten the stance limb. Once again, this sounds plausible until the timing is examined: Stance knee flexion occurs too early, during double support, to provide the necessary shortening, and the knee has already extended again by the time that the COM trajectory is highest in midstance (about 30% cycle).
Ankle rockers: The plantarflexion at toe-off was hypothesized to increase the effective length of the trailing limb at initial contact. In addition, heel-rise during the trailing support phase may be a much more important source of lift during double support.
Rotation of leg segments: Internal rotation and pronation tend to shorten the limb during stance, while external rotation and supination lengthen it during toe-off.
Physiological genu valgum: The normal slight abduction of the knee (genu valgum) was claimed to reduce excessive lateral motion of the COM, bringing the feet closer together and thus reducing the walking base or step width.
Despite being a foundation of GA for many years, the determinants of gait now have been thoroughly discredited. Problems have arisen not only with the timing of many of the mechanisms supposedly aimed at reducing COM displacement, but also the premise that a larger COM excursion is necessarily deleterious to energy consumption. A number of studies have challenged some of these conclusions. Gard and Childress (1997, 1999) investigated the effects of pelvic list and stance knee flexion on the vertical displacement of the trunk and concluded that neither mechanism significantly reduces trunk vertical displacement.69,70 Using a circular rocker foot to simulate the stance phase of gait, these researchers found that anteroposterior translation of the center of pressure (COP) flattens the trajectory of the trunk in the sagittal plane.71