General principles 39
3.1.1 Structural diagnosis 39
3.1.2 Functional diagnosis of spinal column mobility (kinematics) 40
3.1.3 Functional diagnosis of body statics 40
3.2 Technique in functional diagnosis 40
Manual techniques call for an accurate understanding of anatomy, especially when treating the spinal column. Textbooks of radiology are interested mainly in morphology, while our concern is mainly with function, the consistent focus of the present book. The use of radiology for functional studies can greatly improve our understanding in manual diagnosis. However, if we are to be able to interpret radiographic images from the functional point of view, we need a good knowledge of X-ray anatomy, such as we present here, as well as certain requirements as regards technique.
3.1. General principles
For our purposes, X-ray diagnosis fulfills three basic tasks:
1. Structural diagnosis.
2. Functional diagnosis of spinal column mobility (kinematics).
3. Functional diagnosis of body statics (interpretation of curvatures of the spinal column).
3.1.1. Structural diagnosis
Structural diagnosis provides information about the morphology of the bony structures. This is the essential focus of classic X-ray diagnosis, which is mainly concerned with form and structure and constitutes the basis of our knowledge. This type of diagnosis is very important for manual therapy in that it alerts us to potential serious errors of diagnosis, providing a warning against manual therapy where inflammation, tumors, fractures, or other contraindications are present. It also reveals abnormalities and changes in structure, such as asymmetries, which can be significant for function. Diagnosis of structure can be found in the classic textbooks of radiology, so we shall deal here only with those morphological changes that are important for an understanding of dysfunctions.
3.1.2. Functional diagnosis of spinal column mobility (kinematics)
Functional diagnosis in the narrower sense involves movement studies of the spinal column in which X-rays are taken in end of range positions, in ante- and retroflexion (extension), side-bending and, less frequently, rotation. Examination of this kind is the only approach that can provide direct information about dysfunctions in the motion segment. It can also be used before and after treatment. It is of value for documentation and assessment, but is too time-consuming and uneconomical and the radiation exposure is too great for use as a routine procedure. Since manual therapy examination gives good information on mobility and disturbances of mobility, it is generally possible to dispense with movement studies. They do have an important role to play in research, however, and provide an understanding of the biomechanics of movement processes.
3.1.3. Functional diagnosis of body statics
Although movement studies come first to mind when we think of functional diagnosis, it is no less important to diagnose disturbances of body statics. The images used to assess this must be taken standing (or, for the cervical spine, sitting), under static loading and under standard conditions. The curvatures of the spine, as explained below, should mainly be assessed from the point of view of static function. This applies not only to the sagittal but also the frontal plane, in which every obliquity (e.g. of the pelvis during walking) produces a corresponding scoliotic curvature and rotation. Curvature may be regular or irregular, so that a marked deviation may be observed in one particular segment. This may be scoliotic, increased lordosis or kyphosis, rotation, or lateral shift (‘offset’).
The significance of these signs of irregularities in the position of neighboring vertebrae (relational diagnosis) is highly controversial, and closely connected with the discredited subluxation theory. It is also closely linked to the problem of asymmetry, bearing in mind the fact that a degree of asymmetry is the rule rather than the exception. Jirout (1978) has shown that asymmetry of the position of the atlas in relation to the axis is present in the majority of adults. In a study to compare children of various ages, he found that its incidence increases with age. This can be shown particularly easily by observing the position of the spinous processes. He concluded that these asymmetries were the result of asymmetrical pull of the muscles due to the dominance of one cerebral hemisphere.
From this we can conclude that asymmetry and other kinds of irregularity are not in themselves pathological, although they can be the expression of functional asymmetries. We know, for example, that if the axis is asymmetrically rotated in neutral position, the entire cervical spine will rotate asymmetrically during side-bending. In general it is advisable to be cautious when drawing conclusions about examples of asymmetry observed on an X-ray film, and always to take the clinical findings into account when interpreting the radiographic findings.
One advantage of static functional diagnosis is that the examination is economical: only two X-rays are required, two projections, which must correspond to each other vertically. Standard conditions must be observed as regards static loading. As individual posture is highly characteristic, it also remains fairly constant. Gutmann & Véle (1978) said of static function: ‘The dominating principle of the spinal column is body statics. All other functions are subordinate to the requirements of upright posture on two legs. The human body is more ready to accept loss of mobility or painful impingement of nerve roots than to sacrifice erect posture.’
3.2. Technique in functional diagnosis
Since the spinal column is a functional unit, the most appropriate format for the X-ray examination is to show the entire spinal column on a single film. An AP and a lateral view with the patient standing are required, with the feet placed in a standardized position. If this cannot be done, the sections of the spinal column that have been imaged need to be assessed in the light of the clinical findings. These can then make good whatever is missing in the X-ray.
3.3. The lumbar spine and pelvis
3.3.1. X-ray of the lumbar spine and the pelvis
The imaging projections needed for routine examination of the static function and morphological changes of the spinal column are simply one AP and one lateral view with the patient standing. This is done using a device described by Gutmann (1970), in which a plumb line indicates the vertical line from the head. The procedure is illustrated in Figure 3.1 A–D and is as follows:
 |
| Figure 3.1
X-ray technique for the lumbar spine after Gutmann (1970). (A) Positioning of the plumb line for the head and (B) patient position during radiography, as prepared for the AP view. (C) Positioning of the plumb line for the head and (D) patient position during radiography, as prepared for the lateral view.
|
A line which corresponds to the center of the cassette is drawn on the floor in front of the middle of the stand. For the AP view the patient places one foot symmetrically on each side of the line, and is requested to distribute the weight of standing equally between both feet, with legs straight, so as to rest on a base that is in line with the center of the cassette. A plumb line extended downward from the center of the cassette will therefore meet the floor at the mid point between the patient’s heels, setting the base line. A movable plumb line of metal wire (so as to create contrast) is attached to the screen. The cassette is first raised to the level of the patient’s occiput and this metal wire plumb line moved to a point precisely below the middle of the occipital squama, where the external occipital protuberance can be palpated. This sets the plumb line that marks the head position. The cassette is then adjusted (taking care not to displace it to the side) to the height required to take a view of the lumbar region and the pelvis (with the central ray of the X-ray beam and the center of the cassette roughly at the height of the navel). By setting up the AP view in this way, the shadow of the metal wire marks the plumb line representing the head position, and the center of the film represents that of the base line.
The same procedure is used for the lateral view of the lumbar spine, except that this time the patient stands with feet across the line on the floor that represents the center of the cassette, with ankles one finger’s breadth behind the line. The plumb line for the head is positioned at a point in line with the external acoustic meatus. Here – as in the procedure described for the cervical spine – it is helpful to align the central ray below rather than on the center of the cassette, directing the beam off-center to focus approximately on the lumbosacral junction, midway between the iliac crest and the greater trochanter. This technique has two great advantages:
1. There is considerably greater absorption of radiation at the lumbosacral junction (on which the pelvis is superimposed) than in the lumbar spine. If the central ray is aligned as usual on the middle of the lumbar spine, the result is either under-exposure of the lumbosacral junction, while the rest of the lumbar spine is correctly exposed, or over-exposure of the lumbar spine while exposure of the lumbosacral junction is good. Aligning the beam on the lumbosacral junction evens out the exposure, and has the additional advantage of providing good imaging of the hip joints (see Figure 3.2).
2. If the beam is aligned to the center of the cassette (middle of the lumbar spine), the effect is to project the iliac crests away from each other, lying well apart as they do, although they are very important for body statics. The projection error produced at the much slimmer lumbar spine, on the other hand, is only slight.
For both projections the focus-film distance should be as great as possible, depending on the power of the apparatus and the corpulence of the patient, the ideal distance being at least 2m (78inches). For technical reasons, the patient needs to stand with arms folded in front of the chest (see Figure 3.1 D). The final step, once the apparatus and the patient’s position have been set up, is to tape the wire to the cassette so that it cannot be displaced, and then instruct the patient to lean against the screen so as to remain steady during the long film exposure.
3.3.2. X-ray evaluation of lumbar spinal statics
The purpose of the films taken in the standing position is mainly to study body statics. The only structures in the frontal plane that can be assessed by clinical examination are the occipital protuberance, spinous processes, iliac crests, intergluteal cleft, and the midpoint between the heels. In the sagittal plane, clinical examination can show the posture of the head, position of the shoulders, the trochanters and the heels in relation to the plumb line, which takes the line from a fixed point at the external auditory meatus. Clinical examination cannot provide information about the position and inclination of the sacrum and the most caudal vertebrae (which mark the true base of the spinal column), information which is essential for a full understanding and evaluation of spinal statics.
This may explain why clinicians interested in body statics have devoted their attention mainly to the question of body equilibrium as a whole, studying deviation of the head and deviation from the line of gravity by means of statovectography. However, Rash & Burke (1971) pointed out that with static load each body segment should be vertically above the center of the segment on which it rests. The principle is violated if tension of the ligaments or excessive muscular contraction are required to maintain balance. X-ray examination under static conditions provides information on precisely this type of static disturbance.
The mechanism of body statics differs considerably in the frontal and the sagittal planes. One way to appreciate this is to observe the effect of a heel insert, a pad or block placed under one foot. A healthy subject experiences a height difference of as little as 1cm as uncomfortable, whereas if a heel insert of 1cm is placed under both feet it is hardly noticed. This is because in the frontal plane the center of gravity lies over both feet, such that body equilibrium is (relatively) stable. As a result, any mechanical change (the insert under the foot) has an immediate effect. In the sagittal plane, in contrast, body equilibrium is labile, over the two perfectly round surfaces of the hip joints. A slight mechanical change has little effect, because it is dynamic muscle function that is maintaining balance in this plane. The muscular force required should, however, be minimal.
Lumbar spinal statics in the frontal plane
In the ‘ideal’ case the pelvis and spinal column lie symmetrically in a straight line in the AP view. The external occipital protuberance, spinous processes, pubic symphysis, and coccyx lie in the midline. Such a spinal column is the exception in real life; people simply do not place their weight symmetrically on both feet, but stand in a relaxed posture in which the weight is taken mainly on one leg. During walking, the pelvis constantly swings from one side to the other. The result is the constant creation of oblique planes. The main concern in assessing these is to discover how the spinal column reacts to obliquity in the frontal plane.
The physiological reaction to obliquity can be seen by performing a test in healthy subjects: the subject must first relax, stand with legs straight, and rest the weight of the body on both feet; a block of wood is then placed under one foot. The subject’s pelvis then shifts to the higher side (see Figure 3.3).
The radiographic image shows not only the shift to the side, but also scoliosis and rotation to the lower side. The summit of the scoliotic curve is usually at the mid-lumbar region, with the thoracolumbar junction vertically above the sacrum. The degree of rotation depends on the degree of lordosis of the lumbar spine. If there is no lordosis – as is often the case in acute lumbago – there is also no rotation. If there is kyphosis, there may even be rotation to the same side as the concavity.
Reaction by the spinal column to an oblique plane is normal if:
• scoliosis to the lower side results
• there is rotation to the same side (when there is lordosis)
• the thoracolumbar junction is vertically above the sacrum
• the pelvis shifts to the higher side (see Figure 3.4).
Slight thoracic scoliosis occurs in the opposite direction.
These features reflect normal spinal statics and are closely associated with the problem of difference in leg length. From the point of view of body statics, a difference in leg length becomes significant only when accompanied by obliquity of the base of the spinal column (see Figure 3.5).
In the light of this fact, the age-old dispute over how to measure a difference in leg length is beside the point. While it is possible clinically to establish the presence of pelvic tilt, we cannot determine the position of the sacrum relative to the sacral promontory and the lumbar vertebrae that constitute the base of the spinal column proper, as the pelvis may be straight while the sacrum is tilted, or vice versa. The determining factor in spinal column and body statics is the base of the spinal column. The only way to establish the position of these, and find how the spinal column reacts to an oblique base, is by X-ray examination with the patient standing (see Figure 3.6).
The most important findings where there is a disturbance of body statics are:
• obliquity of the base of the spine without scoliosis or with inadequate scoliosis, so that the thoracolumbar junction is not vertically above the lumbosacral
• no lateral pelvic shift to the higher side
• no rotation when there is scoliosis together with lordotic posture of the lumbar spine or even rotation in the direction of the concavity.
The practical decision to be made is whether to order a corrective heel insert. This is primarily a clinical decision, although the X-ray can provide useful clues. The following radiological criteria show when a heel insert can be helpful in the case of obliquity of the base of the spine:
• If scoliosis is not sufficient to bring the thoracolumbar junction into a position vertically above the lumbosacral, or if scoliosis is absent. Use of a heel insert to raise one heel should bring the thoracolumbar junction to the vertical, or at least nearly so.
• If the pelvis is shifted, usually toward the higher side, it should then return to the midline.
• If the scoliosis was statically balanced, it should decrease.
These criteria should all be checked again by X-ray. The spinal column may react positively or negatively to the heel insert, either ‘accepting’ or ‘rejecting’ the correction. If the response is negative it would be wrong to force correction upon the patient, because this would only worsen the situation at the base (see Figure 3.7).
The typical reaction to obliquity as seen radiographically has been studied by Illi (1954) and Edinger & Biedermann (1957), with the subject walking on the spot. At each step, oblique planes appeared together with corresponding scoliosis to the side concerned; the summit of the scoliotic curve appeared at L3. The thoracolumbar junction remained vertically above the sacrum. Above T12 there was a scoliosis of the thoracic spine to the opposite side, but it was shallow. According to Edinger & Biedermann (1957), the thoracolumbar junction forms a kind of point of interchange, and should not swing more than 4cm from one side to the other.
The relation of the scoliosis to rotation and its dependence on the presence of curvature in the sagittal plane was studied by Lovett (1907), who found that rotation of the lumbar spine (in the sense of scoliosis) occurs if lordosis is present, but not in kyphosis. The explanation for this lies in the fact that, whereas the vertebral bodies have good mobility during side-bending, the joints of the vertebral arches are forced together in lordosis and so resist movement. In contrast, in kyphosis there is less close contact between the joints of the vertebral arches; but the vertebral bodies are pressed more firmly against each other, so that these are less free to side-bend. As a result, either there is no rotation at all or rotation occurs in the opposite direction. This is sometimes the case in patients with acute lumbago or in radicular compression syndrome (see Figure 3.8). The situation can also be found clinically in healthy subjects. On passive side-bending of a subject in lordosis, the spinous processes remain in the midline, and the vertebrae rotate in the sense of scoliosis. If the same is done in kyphosis, the spinous processes form a scoliotic curve: in other words, they move in parallel with the vertebral bodies.
Lumbar spinal statics in the sagittal plane
In the sagittal plane, we often speak of ‘normal’ curvatures; these are generally held to be the convex cervical curve (lordosis), concave thoracic curve (kyphosis), convex lumbar curve (lordosis), and concave sacral curve (kyphosis). Sollmann & Breitenbach (1961) demonstrated on the basis of 1000 X-ray films in the sagittal plane that there is no such thing as a general norm; at best we can speak of an ‘individual norm.’ They do not, however, lay down any criteria for this kind of norm.
Cramer (1958) showed, on the basis of 150 measurements of the lumbar spine with the subject standing, that there is a constant correlation between the tilt of L5 and that of T12, and more important still, that the T12 vertebra lies an average of 4cm dorsally to L5. The results of our own study (Lewit 1973) gave complete confirmation of Cramer’s findings and also showed that the plumb line for the head follows a line down from the external acoustic meatus exactly to the navicular bone. We found that the sacral promontory lay an average of 4mm anterior to this plumb line, and the transverse axis of the hip joints 12mm anterior to the plumb line.
Deviations from this norm indicate a disturbance of body statics as a result of lack of muscle coordination. This is most evident in muscle spasm due to acute lumbago or radicular pain, when there is forward-thrust posture (see Figure 3.9), in which the thoracolumbar junction lies exactly over or ventral to the lumbosacral junction. The reverse is found in ‘flabby’ posture, in which the sacral promontory lies well forward of the plumb line for the head, and T12 lies dorsal to L5 by some distance (see Figure 3.10).
 |
| Figure 3.9 |
‘Flabby’ posture is the expression of imbalance of the muscles of the pelvic girdle; it may be the result of weakened abdominal and gluteal muscles, but equally well of hyperactive hip flexors.
The curvature of the lumbar spine is of course also dependent on pelvic tilt which, in turn, varies according to the ‘type’ of pelvis, as is shown in the following section.
One further point to note is that a slight curvature (a ‘flat’ spine) goes hand in hand with hypermobility and lack of stability, while greater curvature (in both the sagittal and the coronal plane) corresponds to stability and less mobility.
The curvatures of the spine are an expression of static function, and should therefore be interpreted in terms of whether they fulfill this function. In the frontal plane, balance is relatively stable; in the sagittal plane, muscle activity is the determining factor. Curvature of the lumbar spine in the sagittal plane is normal if the thoracolumbar junction is dorsal to the lumbosacral, if there is no forward shifting of the sacral promontory (no more than 8cm in front of the center of the cassette, which is double the average). The position of the thoracolumbar junction vertically above the lumbosacral is also the most important criterion in the frontal plane. If there is obliquity at the base, the normal reaction is scoliosis and rotation of the spinal column (if lordosis is present) and a shift of the pelvis to the higher side.
3.3.3. The pelvis
The pelvis and the spinal column together constitute a functional unity, in which the pelvis serves both as the base of the column and at the same time as the connection with the lower limbs. The pelvis transmits motion from the lower limbs, at the same time acting as a shock-absorber. The spinal column rests on the pelvis much as the mast of a boat rests securely on the firm base of the mast step (Benninghoff, 1944). The sacroiliac joints and the pubic symphysis allow for a degree of mobility with enough spring to act as a buffer while also providing adequate stability.
Pelvic types
The function of the pelvis and its influence on body statics depend largely on its type; we owe the recognition of this relationship to Erdmann & Gutmann (1965). The variability observed here is evidence of the phylogenetic instability of this region; evidence of this variability can be seen in our description of the last lumbar vertebra as a ‘transitional’ vertebra; it is difficult here to speak of any such thing as a ‘norm.’ If the variations are asymmetrical as between the two sides, this results in obliquity at the base of the spinal column, and the effects on spinal statics are considerable. If the variations are symmetrical, this affects the length of the sacrum, which is closely associated with its position and inclination.
Erdmann & Gutmann (1965) distinguish the following pelvic types with respect to the associated mechanism of pathology (see Figure 3.11 and Table 3.1). The authors call these Hohes Assimiliationsbecken (high promontory assimilation pelvis), Normalbecken (normal pelvis) and Überlastungsbecken (overload pelvis):
• High promontory: the ‘assimilation’ type presents a long sacrum and high sacral promontory, with a tendency to hypermobility (see Figure 3.17)
• Normal type: this is of average length, with a tendency to restrictions.
• Low promontory: the ‘overload’ type has a low promontory and marked inclination of the sacrum.
All the points summarized in Table 3.1 should be borne in mind when evaluating X-ray films; it will be seen that the type of pelvis affects the spinal curvatures, the height of the last intervertebral disk, and the shape of the vertebral bodies, and therefore also the mobility of the most caudal motion segments. The assessment to be made when hyperlordosis is found will therefore be different in the case of a high promontory (assimilation) type and in that of a low promontory (overload) type. Similarly, a low L5/S1 intervertebral disk will be differently assessed.
Identification of pelvic type (see Figure 3.11 and Table 3.1) is extremely important for the assessment of dysfunctions, especially in the lumbar and pelvic region.
| High promontory (assimilation) | Normal | Low promontory (overload) | |
|---|---|---|---|
| Inclination of sacrum | 50°–70° | 35°–50° | 15°–30° |
| Inclination of end-plate of S1 | 15°–30° | 30°–50° | 50°–70° |
| Position of L4 disk | Above the line of the iliac crests | At the height of the iliac crests | Below the line of the iliac crests |
| Position of the promontory in the pelvic girdle | Eccentric (dorsal) | At the center | At the center or ventral |
| Shape of L5 vertebra | Rectangular | Trapeze shaped | Trapeze shaped |
| Shape of L5 disk | Rectangular and higher than L4 | Wedge shaped and thinner than L4 | Wedge shaped and thinner than L4 |
| Segment with maximum mobility | L5/S1 | L4/L5 | L4/L5 |
| Effect of iliolumbar ligament | Slight fixation of L5 | Good fixation of L5 | Good fixation of L4 and L5 |
| Weight-bearing structure | End-plate of S1 | End-plate of S1 | L5/S1 joints and sacroiliac joint |
| Spinal curvature | Flat | Average | Considerable |
| X-ray statics | Promontory and hip joints lie in front of plumb line for head | Promontory and hip joints almost on line of plumb line for head | Promontory and hip joints lie behind plumb line for head |
| Clinical signs | Hypermobility, pathological changes of L5 disk; ligament pain | Restrictions, pathological changes of L4 disk | Arthroses of L5/S1, sacroiliac joint and hip |
The sacroiliac joints
There is some mobility of the otherwise firm pelvic girdle, due to the role of the sacroiliac joints and the pubic symphysis. The major role is that of the sacroiliac joints.
The sacrum is wedge shaped in two directions; first the whole structure tapers in the caudal direction. A double contour is usually seen in the AP view, since there is another wedge in the ventrodorsal direction; the sacrum is somewhat broader ventrally, at least in its craniad part, although in this respect too there are considerable variations. It is helpful to note that the greater the distance between the two contours of the joint, the narrower the joint space appears. If on the other hand we see only one contour, the joint space appears to be wide and clearly defined. This is often the case with the high promontory type, and is a further sign of hypermobility.
It is important to point out that, despite its unusual shape and limited mobility and the fact that there are no muscles to move the sacrum against the ilium, the sacroiliac joint is a true synovial joint (Colachis et al 1963; Duckworth 1970; Mennell 1952; Weisl 1954). According to Duckworth, the sacrum rotates relative to the ilia around an axis corresponding to the shortest sacroiliac ligaments, at the level of S2. This movement is one of nutation; with each step taken during walking, the weight of the spinal column produces a forward nodding motion of the sacrum, together with the sacral promontory, acting as a shock-absorber. This mobility of the sacrum within the pelvic girdle is easily palpated and is familiar to gynecologists in the management of labor. Perpendicular to this ‘functional’ motion, the joint play consists in a springing, wing-like motion about a craniocaudal axis, the effect of which is a distraction of the joint.
The degree of mobility in the sacroiliac joint should be as little as possible, yet never to the point of restriction, just as a shock-absorber is firm but never immobile.
It is appropriate at this point to deal with a condition described as pelvic distortion, which requires explanation from the functional anatomical point of view. The finding on palpation is that the posterior superior iliac spine (PSIS) is lower on one side than the other. The finding is the same if the finding is made at the posterior border of the iliac crests at the point where they can be palpated in the vertebral region. Ventrally the opposite is found: on the side where the PSIS is lower, the anterior superior iliac spine (ASIS) is
found to be higher than on the contralateral side, and vice versa. The ventral parts of the iliac crests behave in the same way as the anterior iliac spines. The middle portion of the iliac crests may be symmetrical, although this need not be so. On first impression it seems as if one ilium is twisted relative to the other about a frontal transverse axis, although this is in fact impossible if the pubic symphysis is intact.The functional anatomy involved can best be illustrated anatomically by Cramer’s diagram (1965) (Figure 3.12). This shows a one-sided nutation of the sacrum brought about by its rotation between the ilia around its longitudinal axis. This in turn results in rotation of one ilium about a horizontal axis and of the other about a vertical one.
 |
| Figure 3.12 |
All attempts to visualize these changes radiographically have remained without success as far as we know. However, we have been successful in showing X-ray evidence of a disturbance of body statics in the presence of pelvic distortion (see Figure 3.13). The pelvis was found to be shifted toward the higher side, and there was deviation of the angle between the sacrum and the lumbar spine. This disappeared following treatment of the atlanto-occipital and atlantoaxial joints. Lewit & Rosina (1999) were able to induce pelvic distortion by rotating the head to one side and then the other, but radiographic examination showed this effect to have been a palpatory illusion.
3.3.4. The lumbar spine
Although only a little shorter than the thoracic spine, the lumbar spine consists of only five vertebrae. However, in ante- and retroflexion as well as in side-bending, the corresponding motion segments play a significant part in ensuring the mobility of the trunk. Meanwhile the inferior part of the lumbar spine also carries the weight of the trunk, so the vertebral bodies and articular processes of the lumbar spine are the most robust.
The joints of the vertebral arches form massive gliding surfaces that can enable considerable excursion and also provide stability. The greater part of the articular facets runs vertically, almost in the sagittal plane. Ventrally, the smaller part is turned almost at a right angle to point medially, in the frontal plane. Frequently, however, the articular facets simply form an arc, whose concave aspect faces dorsally. If the two parts do stand at right angles to each other, the joint spaces can be easily visualized by X-ray; if they form an arc, this cannot be done. Given that the joints of the vertebral arches only develop their final shape after birth, during the first years of life there is considerable variation in this respect.
Buy Membership for Rheumatology Category to continue reading. Learn more here
