Reliability of ultrasonographic measurement of vertebral artery blood flow.
Background: Clinical tests involving sustained cervical spine rotation and/or extension are commonly applied pre-manipulatively to screen for patients at risk of stroke due to vertebral artery pathology. This is despite the fact that the validity of these manoeuvres is disputed and their effect on vertebral artery blood flow poorly understood. Recent research has employed duplex ultrasound (Doppler with B-mode real-time imaging capability) to quantify positional haemodynamic parameters and determine whether subjects testing positive differ from negative subjects. However, the reliability of Doppler sampling of the upper cervical part of the vertebral artery in positions involving rotation and extension has not been established. Thus the aim of this study was to evaluate the reliability and measurement variability associated with this investigative procedure.
Methods: Twenty normal subjects volunteered to participate in the study. Haemodynamic measurements were taken of a randomly selected vertebral artery using duplex ultrasound with colour flow and power Doppler imaging capabilities. Blood flow was recorded at both the atlanto-axial and the C2/3 regions of the vessel in neutral, end-range extension and end-range contralateral rotation. The protocol was then repeated.
Results: Intraclass correlation coefficients (2, 1) and 95% limits of agreement indicated that sampling at the more vulnerable but less accessible atlanto-axial site was generally more repeatable, notably in end-range contralateral rotation. In particular, the key measures of atlanto-axial peak systolic velocity and resistance index in end-range rotation demonstrated excellent reliability (0.82; 0.76). Between group differences of -12.9 to 16.6 cm/s and -0.11 to 0.17 respectively would be necessary to discount measurement variability.
Conclusions: Accurate interpretation of the results of ultrasonographic investigation of vertebral artery tests utilising rotation and extension requires consideration of measurement variability and reliability. Haemodynamic parameters of acceptable reliability and associated ranges of measurement variability have been identified for use in future research.
Key Words: Manipulation Therapy; Vertebral Artery; Doppler Ultrasonography; Reproducibility of Results
Cervical spine manipulation (CSM) is an intervention commonly utilised by physiotherapists, medical practitioners, chiropractors, and osteopaths for the alleviation of local and referred cervicogenic pain. The relief reportedly attainable from CSM has contributed to an increasing acceptance of this form of treatment, with an estimated 18 to 38 million cervical spine manipulations performed annually just in the United States by chiropractors alone Shekelle and Coulter, 1997).
Although the cervical spine has not received as much attention in the clinical research literature as has the lumbar spine, evidence supporting the efficacy of CSM is gradually growing. Whilst acknowledging the need for further research, the evidence suggests that CSM is an effective treatment for some patients with subacute or chronic neck pain (Hurwitz et al, 1996; Shekelle and Coulter, 1997). In addition, there is some support for its use in the management of cervicogenic headache (Nilsson et al, 1997), as well as muscle tension-type (Boline et al, 1995; Hurwitz et al, 1996) and migraine headache (Parker et al, 1978; Nelson et al, 1998). In the latter two types of headache there is evidence to suggest that it may be as clinically effective as amitriptyline but with relatively fewer side effects (Boline et al, 1995; Nelson et al, 1998). However, despite CSM being arguably safer than some alternate therapies such as nonsteroidal anti-inflammatory drugs (Dabbs and Lauretti, 1995; Rivett, 1995), much controversy still surrounds its causal relationship with extracranial arterial trauma and consequent stroke (Adams and Sim, 1998: Rivett and Reid, 1998).
While the internal carotid artery has been implicated in a small proportion of these incidents (Lee et al, 1995; Rivett and Milburn, 1997), it is accepted in the literature that the vertebral artery (VA) is most vulnerable, particularly the atlanto-axial part during cervical spine rotation (Terrett, 1987; Grant, 1994; Refshauge, 1995; Kuether et al, 1997). As approximately half of overall cervical spine rotation occurs at the C1/2 level, the contralateral vessel is subjected to stretching during this movement (Terrett, 1987; Penning, 1988; Grant, 1994). Case study descriptions have provided confirmatory evidence of the susceptibility of the atlanto-axial region of the contralateral VA to rotatory manipulative trauma (Daneshmend et al, 1984; Sinel and Smith, 1993; Grant, 1994; Hurwitz et al, 1996). Recently Rivett and Reid (1998) estimated that the risk of stroke following manipulation applied by a physiotherapist to be about one in every 163,000 neck manipulations. Although it is generally agreed that serious neurovascular complications are rare events, the potential for permanent disability or even death has led to the development of various pre-manipulative tests and clinical protocols designed to detect the patient at risk of manipulative stroke (APA, 1988; Aspinall, 1989; Lewit, 1992; Ivancic et al, 1993; Carey, 1995). Usually these tests involve sustained neck rotation and/or extension in order to provoke signs and/or symptoms indicative of vertebrobasilar insufficiency. However the validity of the recommended pre-manipulative tests has not been established and has become a topic of some controversy in itself (Refshauge, 1994; Thiel et al, 1994; Cote et al, 1996; Rivett et al, 1999).
Doppler ultrasound has become the primary investigative tool used to evaluate in vivo the validity of pre-manipulative tests primarily by determining the effects of cervical spine rotation and/or extension on various VA blood flow parameters. Diagnostic ultrasound has many attributes that make it an attractive research tool: it is non-invasive, relatively inexpensive, completely safe and generally accessible. It has also been shown to be a valid and reliable instrument for the standard measurement of extracranial arterial blood flow (Ringelstein et al, 1985; Lo et al, 1986; Karnik et al, 1987; Colquhoun et al, 1992; Fry et al, 1994; Young et al, 1994). Nevertheless there has been disagreement in the literature as to the effects of various neck positions on VA flow, with some researchers finding significant flow change (most notably during rotation) (Stevens, 1991; Refshauge, 1994; Licht et al, 1998; Rivett et al, 1999), whereas others have produced conflicting results (Weingart and Bischoff, 1992; Thiel et al, 1994). Furthermore, Thiel et al (1994) and Cote et al (1996) have disputed the validity of pre-manipulative tests based on different analyses of data from the same duplex ultrasound study of the effects of positional tests on both positive and negative subjects. Recently Rivett et al (1999) conducted a pilot study comparing ten subjects testing positive to ten subjects testing negative. Significant changes in VA flow velocity at the C2/3 level were shown for both groups in end-range positions involving rotation and extension using duplex ultrasound, although the pilot study was too small to examine differences between the two groups. The authors are currently carrying out a larger study to determine whether or not subjects testing positive significantly differ from those testing negative.
There may be several methodological reasons why there is a lack of consensus as to the haemodynamic effects and the validity of pre-manipulative tests. Some earlier investigations (Stevens, 1991; Weingart and Bischoff, 1992) utilised continuous-wave Doppler without imaging capabilities, which raises the possibility of inaccurate sampling (Licht et al, 1998). Error may result simply from missing the target vessel (or perhaps sampling another vessel) and also from unknown changes in the angle of insonation. Secondly, some studies have used only normal volunteers which limits the generalisability of the findings to the patient population and hence what can be inferred regarding the validity of the tests (Weingart and Bischoff, 1992; Refshauge, 1994; Licht et al, 1998).
Furthermore, few investigations have actually attempted to sample at the vulnerable atlanto-axial region of the VA (Stevens, 1991; Haynes, 1996). However, it is arguably more desirable to sample at this region of the vessel as it is more likely to undergo diameter and blood flow velocity changes resulting from rotation manoeuvres (Grant and Johnson, 1997; Licht et al, 1998). Presumably the unique high cervical spine bony architecture coupled with the tortuosity of the VA and prevalence of vascular anomalies in this region has rendered more caudal sites easier targets. Recently a new sonographic technique referred to as power Doppler imaging (PDI) has become available and has several features that potentially facilitate the visualising of continuity of blood flow in the atlanto-axial region of the VA. PDI provides colour imaging of vessels independent of the angle of insonation and independent of flow velocity and direction as it relies only on the density of red blood cell aggregation (Griewing et al, 1996; Steinke et al, 1997). These features may assist in tracing the winding course of the VA in the high cervical spine and in instances of cessation of flow, such as with mechanical vessel occlusion. Ideally it is used in concert with colour Doppler flow imaging which provides a colour-coded visualisation of the velocity and flow direction of moving blood cells, which can thus specifically help to detect areas of abnormal haemodynamics (such as retrograde flow) and differentiate arterial from venous structures (Griewing et al, 1996; Steinke et al, 1996; Steinke et al, 1997).
Finally, most studies have failed to address the issue of the reliability or repeatability of the ultrasonographic measures. This is despite the fact that reliability is of the utmost importance in ultrasonographic investigations as the accuracy of the measurements is largely operator dependent (Mikkonen et al, 1996; Grant and Johnson, 1997). Stevens (1991) states that he obtained consistent measurements on three occasions for each of a variety of test positions but does not provide any supportive data or analysis. Refshauge (1994) reported that recordings of peak and mean frequency were highly reproducible for measurements taken at the C2/3 level of the VA. However the figures given for absolute intertest differences were indicative of overall repeatability across three positions (neutral; 45[degrees] contralateral rotation; full range contralateral rotation) and were therefore not specific for any one position. More recently Grant and Johnson (1997) evaluated the reliability of sampling in the neutral position using a range of haemodynamic parameters at a variety of sites, including at the C1/2 level. An intraclass correlation coefficient (ICC) of 0.19 or less was found for averaging up to five measurements of peak velocity at the atlanto-axial region, and despite this being indicative of poor reliability, the authors argue that the likely degree of change found on rotation would be greater than the measurement variability. Caution is required in extrapolating the results of this study to sampling undertaken in other positions, most notably end-range rotation, due to effects of the biomechanics of the high cervical spine on the relationships of bony and vascular structures and consequently the haemodynamics of the VA.
Therefore, the primary aim of the present study was to evaluate and compare the reliability and measurement variability of sampling at both the atlanto-axial and C2/3 regions of the VA in neutral, end-range extension and end-range contralateral rotation, using a variety of haemodynamic parameters. A further aim was to assess the feasibility of using PDI in tandem with colour flow imaging to visualise the atlanto-axial part of the VA in the various positions. The results of this study will assist in investigating whether subjects found positive on premanipulative testing differ from negative subjects.
Twenty subjects volunteered to participate in the study in response to an advertisement; twelve were female and eight were male, and the mean age was 35.5 years (SD = 9.3, range 24-55). An information sheet was provided to each subject and any questions were answered, following which written informed consent was obtained. Exclusion criteria included disorders (eg previous stroke) and medications (eg anticoagulant medication) that would normally contraindicate CSM (Grieve, 1989; Kenna and Murtagh, 1989) and that may have put subjects at risk of injury. Potential participants were also excluded if they currently suffered from marked postural hypotension or inner ear conditions. In addition, subjects were only admitted to the study if they were aged between 18 to 70 years. Subjects were also matched for age and gender to the group investigated in an earlier pilot study conducted by the authors (Rivett et al, 1999). The subjects in the pilot study were predominantly referred to the investigators by community physiotherapists who had previously performed pre-manipulative testing of these patients in the normal course of their clinical practice. Therefore, findings of the present study could be generalised to the patient population undergoing pre-manipulative testing.
Following admission to the study each subject underwent an ultrasonographic investigation of one randomly selected vertebral artery in a radiological clinic. The resultant sample of twenty VAs was equally divided between left and right vessels. Vascular measurements were taken using a 2D Gateway Series duplex ultrasound apparatus with B-mode real-time imaging, power Doppler imaging (Angio), colour flow mapping capability, and a 5MHz linear transducer (Diasonics Ultrasound, Santa Clara, CA). All scanning was conducted by a single qualified ultrasonographer with extensive experience in the examination of the extra-cranial vasculature and who had performed all examinations in the preceding pilot study (Rivett et al, 1999). Several practice sessions were undertaken prior to data collection to ensure that the ultrasonographer was familiar with the study protocol.
The procedure commenced with a ten minute rest period to assist in facilitating haemodynamic stability. A Digital Blood Pressure Monitor model DS-145 (ALPK2, Japan) was used to record the subject's blood pressure (BP) and pulse rate (PR) at the conclusion of this period with an accuracy of within [+ or -]3mmHg for BP and [+ or -]5% for PR. Initial ultrasonographic measurements of the VA were taken with the cervical spine in a neutral position. Measures recorded included: peak systolic velocity; end diastolic velocity; time averaged velocity; flow rate; lumen diameter; and two indices of vascular impedance, the Pourcelot or resistance index and the systolic/diastolic ratio (McDicken, 1991). The diameter of the lumen was measured perpendicular to the long axis of the vessel using the sample volume size indicators. Recordings were then made whilst the cervical spine was held firstly in the position of end-range extension and then in endrange contralateral rotation. These positions were chosen as they constituted test positions commonly applied pre-manipulatively, both individually and in combination. In addition, contralateral rotation was investigated rather than ipsilateral rotation as this movement is more frequently cited as being associated with iatrogenic trauma and with changes in blood flow (Daneshmend et al, 1984; Grant, 1994; Kunnasmaa and Thiel, 1994; Rivett et al, 1998).
In order to avoid inaccuracies of measuring blood flow velocity related to beam-to-vessel angle, the angle of insonation was maintained at 60 degrees or less to the artery during recording (Nelson and Pretorius, 1988). Measurements of the vessel were performed firstly in the region of the artery between the second and third cervical vertebrae and then between the first and second vertebrae in each position. The use of PDI with its angiographic-like image enabled the ultrasonographer to visualise the continuity of the VA blood flow in the tortuous atlanto-axial region independent of the angle of insonation and blood flow velocity. The characteristic curvature of the VA in this region was used to accurately identify the location for sampling. The ultrasonographer switched to colour Doppler flow imaging in order to confirm flow direction and velocity and to take measurements. Upon any flow changes being noted during testing the contact pressure of the transducer was altered so as to ensure that local pressure was not contributing. Scans were conducted with all subjects lying supine.
In the event of a subject reporting any marked discomfort in an end-range position then the position was reduced to that point in the range of motion just short of the onset of the discomfort. Test positions were sustained for at least 30 seconds unless there was provocation of possible ischaemic symptoms or pain, at which time measurements were immediately recorded and the neck returned to neutral. This period was selected as it permitted the ultrasonographer adequate time to accurately locate the target site without maintaining the potentially stressful endrange position for an unnecessarily long duration. Following each test position the cervical spine was returned to the neutral position for 30 seconds in order to facilitate the recovery of blood flow parameters to baseline values and to wait for any latent symptoms.
The ranges of motion for both extension and rotation were measured using a modified Cervical Range-of-Motion (CROM) instrument (Performance Attainment Associates, St. Paul, MN, USA). This tool has been shown to be highly reliable in measuring the range of motion of these movements (Youdas et al, 1991; Rheault et al, 1992). To permit use in supine lying the sagittal plane inclinometer was re-oriented and the compass was reversed with the lateral plane inclinometer. The principal investigator (DR) maintained the position of the cervical spine by the application of mild manual pressure to the subject's head and by observing the values from the CROM device.
After the initial scanning in all three positions the ultrasonographic examination was then repeated following the same format. Ranges of motion for extension and rotation were recorded on both occasions of scanning. The subject's BP and PR were again measured at the conclusion of the entire investigation.
A radiologist reviewed all the ultrasound scans for evidence of structural abnormalities of the vasculature and for cardiovascular pathology. The study was approved by the Southern Regional Health Ethics Committee (Otago).
Repeatability of each ultrasound measurement was assessed by determining the 95% limits of agreement derived from the mean and standard deviation of the differences between the two sets of recordings, as described by Bland and Altman (1986). The 95% limits of agreement indicate the range of values within which 95% of the differences between two repeated measurements are expected to lie. In addition, intraclass correlation coefficients (2, 1) with 95% confidence intervals (CI) were calculated. Analyses were carried out for both regions of the vertebral artery, in each of the three positions and for all seven haemodynamic parameters.
Haemodynamic stability was evaluated by comparing systolic BP, diastolic BP and PR measurements taken immediately before and after the ultrasound examination, using mean differences, standard deviation of the differences, and paired two-tailed t tests. The consistency between the two sets of measurements of range of motion for extension and rotation was also assessed using mean differences, standard deviation of the differences, and paired two-tailed t tests. A significance level of a = 0.05 was set for the t tests.
Only one subject reported symptoms during the ultrasound examination that may have possibly been of neurovascular ischaemic origin (transient mild nausea following end-range extension). Another subject had apparent complete cessation of blood flow of the right VA in end-range contralateral rotation yet did not experience any symptoms. This finding was evident at both sites of sampling and on both occasions of measurement (systolic/diastolic ratios were unable to be calculated due to zero velocity values). In addition, the atlanto-axial region of the VA of the first subject could not be found in neutral during the first trial (although it was identified in the second trial), hence this individual was excluded from analysis of C1/2 measurements taken in this position. No abnormalities were found by the radiologist on reviewing the scans of all twenty subjects.
Diastolic BP and PR readings taken just prior to the ultrasound examination were not significantly different to those obtained immediately afterwards (Table 1). However, there was a significant decrease in the mean systolic BP measurements of 8.8mmHg. Range of motion measurements for rotation did not significantly differ, with a difference of the means of less than 1 degree. Although the extension range of motion recordings were found to be significantly different, the actual difference of the mean values was clinically quite small (1.6 degrees).
Reliability of Atlanto-Axial Measurements
The results of the analysis of the measurements taken at the atlanto-axial region of the vertebral artery are presented in Table 2. There is a tendency for the ICC values to be higher for the rotation measurements than for those recorded in neutral and extension, although the limits of agreement tend to be narrowest in neutral. Most ICC values for parameters demonstrated fair to good reliability (0.40 to 0.75), with none showing poor reliability (<0.40) (Fleiss, 1986). Notably, the measurements of resistance index (0.76) and peak systolic velocity (0.82) in end-range contralateral rotation were found to have excellent reliability (>0.75), as were the recordings for flow rate (0.76) in the neutral position and lumen diameter in end-range extension (0.79). The resistance index and peak systolic velocity measurements in end-range contralateral rotation are of particular interest, because this is the position considered to principally stress the vulnerable atlanto-axial region of the vessel during pre-manipulative testing. For the resistance index measure, there is a negligible mean difference indicating that there is minimal bias between the two sets of observations and that they agree excellently on average. The 95% limits of agreement indicate that the differences between repeated measures are unlikely to be more than -0.11 or +0.17. In Figure 1 the differences between each pair of observations were plotted against the average of the two measurements, and the 95% limits of agreement (mean +/- 2SD) are illustrated. The differences do not appear to be related to the size of the resistance index measurement. Figure 2 depicts the relationship between the differences and the averages of the repeated observations for peak systolic velocity in end-range rotation, including the 95% limits of agreement (-12.9, 16.6). Thus repeated measures of peak systolic blood flow are unlikely to differ by more than -12.9 or +16.6 cm/s.
[FIGURES 1-2 OMITTED]
Reliability of C2-C3 Measurements
The results of the analysis of measurements taken between the second and third cervical vertebrae in the various positions are shown in Table 3. The ICC values for recordings made in the neutral and endrange contralateral rotation positions are generally higher than those determined for the vessel parameters in the end-range extension position. The limits of agreement also tend to be narrowest in the neutral position. Again, most of the measures demonstrate fair to good reliability, although a few are categorised as poor: end diastolic velocity (0.22), time averaged velocity (-0.10) and flow rate (0.23) all measured in extension. Two of the parameters demonstrating excellent reliability when measured at the atlanto-axial portion of the artery displayed similar reliability for recordings made from the C2/3 region (peak systolic velocity [0.76] in rotation; lumen diameter [0.84] in extension). In addition, high ICC values were associated with measurements of peak systolic velocity (0.84) and lumen diameter (0.80) in neutral, as well as lumen diameter (0.94) in rotation. The excellent reliability of these parameters is also generally reflected in their negligible mean differences and relatively narrow 95% limits of agreement.
On comparing the findings from the two measurement sites of the vessel there is a tendency for the ICC values to be higher for recordings taken at the atlanto-axial region in the positions of endrange extension and end-range contralateral rotation than for those obtained at the C2/3 region. For values calculated in neutral the converse appears to be true. However, the 95% limits of agreement are generally narrower for measures determined from readings taken at the more cephalad site in both the rotation position and the neutral position. The ICC values obtained from measurements at the atlanto-axial region are also generally superior (or at least equivalent) to those determined for the C2/3 region for all haemodynamic parameters, with the main exception of the lumen diameter. This is also reflected by the limits of agreement which are generally narrower for all parameters measured at C1/2 as opposed to C2/3, apart from the lumen diameter.
The reliability of measuring certain haemodynamic parameters at both the atlanto-axial and C2/3 regions of the VA has been demonstrated to be satisfactory, particularly in neutral and end-range contralateral rotation.
Peak Systolic Velocity Measurement
The measure of peak systolic blood flow velocity recorded in the rotation position at both sites has yielded high ICC (2, 1) values. The adequate repeatability of this parameter is also supported by the small mean differences (indicating negligible bias in the observations) and by the relatively narrow limits of agreement (indicating acceptable measurement variability). Furthermore, the ICC values for peak systolic velocity in the present study are superior to those reported by Grant and Johnson (1997) at both C1/2 (<0.19) and C2/3 (0.37 - 0.63) for single or averaged multiple measurements in the neutral position. Peak systolic velocity has commonly been investigated in previous studies of VA haemodynamics in cervical rotation, suggesting that researchers regard it as a useful measure of blood flow change (Stevens, 1991; Refshauge, 1994; Licht et al, 1998; Rivett et al, 1999). Although these studies report results obtained from a variety of sampling sites they are consistent in their conclusions that contralateral rotation leads to changes in peak systolic blood flow velocity, usually diminishing at end-range.
Resistance Index and End Diastolic Velocity Measurement
Furthermore, the ICC value for the resistance index measured at the atlanto-axial site in end-range contralateral rotation was also excellent (0.76), with a negligible mean difference and relatively narrow 95% limits of agreement (-0.11, 0.17). This suggests that the resistance index may be a useful parameter for measuring differences in blood flow in this position. Nevertheless, Weingart and Bischoff (1992) using a modified Pourcelot index found no significant change in VA flow with any position tested, including endrange rotation. Considering the excellent reliability of the resistance index in end-range rotation shown in the present study using duplex ultrasound (with realtime imaging), measurement error would seem to be an unlikely explanation for their differing results. However Weingart and Bischoff (1992) used continuous-wave Doppler without any imaging capability in sampling at the difficult to access VA region near the arch of the atlas and this may have led to error in identification of the vessel. Notably Rivett et al (1999) found a significant decrease in the resistance index for their negative group using duplex ultrasound of the right VA in end-range rotation and combined rotation/extension. It is also remarkable that the decrease in resistance index was due to a proportionally greater reduction in peak systolic velocity than end diastolic velocity (Stevens, 1991; Rivett et al, 1999). Thus despite generally showing only fair to good reliability in the present study, end diastolic velocity is arguably of value in helping to explain the haemodynamics of the VA in various positions, particularly changes in the various indices of vascular impedance.
Lumen Diameter Measurement
Measurements of lumen diameter at both sites also demonstrated good to excellent ICC values and small mean differences. Furthermore, for all lumen diameter measurements the limits of agreement are less than or equal to one millimetre in either direction indicating that a difference of greater than this amount would likely be due to another factor apart from variability of measurement. However, caution is required in interpreting these results as the sample volume size indicators were used to measure the vessel diameter in each position. The sample volume size indicators measure in increments of minimally 0.3 mm rather than the 0.1mm increments normally used for the B-mode electronic callipers. This was done as part of the process of determining the flow rate and therefore minimised the time subjects were held in each endrange position. The repeatability of the lumen diameter measure when recorded at the second segment of the VA is further supported by the consistent findings of previous studies (Refshauge, 1994; Thiel et al, 1994), albeit that these were measured only in the neutral position. However, Stevens (1991) reports seven cases in which narrowing of the atlanto-axial lumen of the VA was constantly shown in 45. contralateral rotation but fails to provide any data.
Flow Rate Measurement
The ICC values for the flow rate measures vary greatly, from the poor value obtained for the C2/3 sample in end-range extension (0.23) to the excellent reliability of recordings at C1/2 in neutral (0.76). Similarly, the mean differences vary markedly across the flow rate measurements ranging from -11.67 to -0.83 ml/m. However, without exception all flow rate readings displayed wide limits of agreement, indicating marked variability in observations, and thus limiting the usefulness of these measures. This finding is not surprising as the determination of the flow rate is dependent upon the accuracy of both the lumen diameter and average velocity measurements according to the equation:
flow rate = cross-section x average velocity
In particular, an error in lumen diameter measurement will result in compounding of the error in the cross-sectional area calculation as the diameter is squared in the equation, rendering flow rate measurements of doubtful value for small vessels such as the VA (McDicken, 1991). There are no studies using flow rate for the evaluation of positional effects on VA flow suggesting that other investigators may share these concerns. Nevertheless, Grant and Johnson (1997) reported an ICC value of 0.81 for a single flow rate measurement at C5/6 in the neutral position, perhaps reflecting the relative ease of access to the VA at this more caudal site.
Systolic/Diastolic Velocity Ratio Measurement
The systolic/diastolic velocity ratio also provided fair to good ICC values and mostly small mean differences when measured at either site. It is therefore of interest that Thiel et al (1994) used systolic/diastolic ratio findings to conclude that head and neck positions had little effect on VA blood flow. Cote et al (1996) also conducted their analysis using this haemodynamic parameter and came to a similar conclusion. In considering the findings of these two investigations the moderately wide limits of agreement for this parameter found in the present study suggests that the disparity between these two reports and the uniform conclusions of other studies may at least be partly attributable to the variability of this measure, in addition to the small size (n = 12) of their experimental group. Grant and Johnson (1997) also argue that the conclusions of Thiel et al (1994) should be viewed with caution as they found an ICC value of just 0.55 - 0.57 for a single systolic/ diastolic ratio measurement at C5/6 in neutral.
Time Averaged Velocity Measurement
On the other hand, the equivocal results obtained for time averaged velocity are somewhat unexpected as Refshauge (1994) reported that mean frequency measured at C2/3 was highly reproducible in both neutral and also end-range contralateral rotation. Differing findings were however reported by Grant and Johnson (1997) who experienced technical difficulties in sampling at C2/3 and determined an ICC value of only 0.39 (compared to 0.68 for the present study) for time averaged velocity recorded in neutral, although the value improved with averaging multiple measurements. The wide limits of agreement found in the present study for all time averaged velocity measurements would also suggest that this parameter is likely to be of minimal use in detecting differences in blood flow.
The benefits of PDI in helping to visualise the continuity of VA flow in the high cervical spine may have been a contributing factor to the superior repeatability shown by nearly all measures evaluated at the atlanto-axial region in the various positions, including notably end-range rotation. Recording at this cephalad site is particularly desirable as this region of the artery is the most vulnerable to the stresses associated with rotation and is therefore likely to demonstrate the greatest diameter and flow changes (Terrett, 1987; Stevens, 1991; Grant and Johnson, 1997). The finding that measurements at the atlanto-axial site in end-range rotation are of relatively high reliability is somewhat surprising as insonation of the extra-cranial vessels is normally clinically conducted with the neck only in neutral or very slight contralateral rotation. Perhaps the physical positioning of the transducer and hence access to the atlanto-axial loop was facilitated in this position. Nevertheless this is an advantage as rotation is arguably the position of greatest relevance to the pre-manipulative tests.
The differences in range of motion between the two instances of measurement are small and unlikely to greatly contribute to the variability in the scores. Despite the results of the t test indicating a significant difference in the extension range, the actual difference between the means was only 1.6 degrees. In addition, PR and diastolic BP measurements were stable, although the 8.8mmHg mean difference in systolic BP may possibly have been a contributing factor to the variability in the scores. This difference was somewhat unexpected as systolic BP was found to be stable in the pilot study (Rivett et al, 1999).
The results of this study demonstrate that measurements of most haemodynamic parameters in the vulnerable atlanto-axial part of the vertebral artery are generally of higher reliability than those obtained at C2/3, particularly in end-range contralateral rotation. This is despite the technical challenges posed by the unique anatomical structure and the frequency of vascular anomalies in this region of the cervical spine. The use of the recently developed power Doppler imaging, in addition to colour flow imaging, has helped in this regard by permitting relatively easier visualisation of the tortuous high cervical route of the vessel. Notably high ICC values and relatively narrow limits of agreement for peak systolic blood flow velocity and resistance index in end-range contralateral rotation at C1/2 indicate these measures may be of particular use in determining whether subjects testing positive to pre-manipulative testing differ from those testing negative. This recommendation is further supported by the consideration that end-range contralateral rotation of the VA, particularly the atlanto-axial region, is frequently implicated in vertebrobasilar insufficiency and manipulative stroke. The limits of agreement for each parameter also provide a range of values that indicate the difference required to discount measurement variability as the source of the difference. These findings provide a basis upon which further research of VA blood flow using duplex ultrasound can be appropriately designed and accurately interpreted.
The authors wish to thank Leanne Bardwell, ultrasonographer, for conducting the ultrasound examinations and Dr Neil Morrison, radiologist, for helping with interpretation of the scans.
Adams G, Sim J. A survey of UK manual therapists' practice of and attitudes towards manipulation and its complications. Physiotherapy Research International 1998; 3: 206-227.
Aspinall W. Clinical testing for cervical mechanical disorders which produce ischemic vertigo. The Journal of Orthopaedic and Sports Physical Therapy 1989; 11: 176-182.
Australian Physiotherapy Association. Protocol for premanipulative testing of the cervical spine. Australian Journal of Physiotherapy 1988; 34: 97-100.
Bland MJ, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet 1986; 1(8476): 307-310.
Boline PD, Kassak K, Bronfort G, Nelson C, Anderson AV. Spinal manipulation vs. amitriptyline for the treatment of chronic tension-type headaches: a randomized clinical trial. Journal of Manipulative and Physiological Therapeutics 1995; 18: 148-154.
Carey PF. A suggested protocol for the examination and treatment of the cervical spine: managing the risk. The Journal of the Canadian Chiropractic Association 1995; 39: 35-39.
Colquhoun I, Oates CP, Martin K, Hall K, Whittingham TA. The assessment of carotid and vertebral arteries: a comparison of CFM duplex ultrasound with intravenous digital subtraction angiography. The British Journal of Radiology 1992; 65: 1069-1074.
Cote P, Kreitz BG, Cassidy JD, Thiel H. The validity of the extension-rotation test as a clinical screening procedure before neck manipulation: a secondary analysis. Journal of Manipulative and Physiological Therapeutics 1996; 19: 159-164.
Dabbs V, Lauretti WJ. A risk assessment of cervical manipulation vs. NSAIDs for the treatment of neck pain. Journal of Manipulative and Physiological Therapeutics 1995; 18: 530-536.
Daneshmend TK, Hewer RL, Bradshaw JR. Acute brain stem stroke during neck manipulation. British Medical Journal 1984; 288: 189.
Fleiss JL. The Design and Analysis of Clinical Experiments. New York: John Wiley and Sons, 1986; 7.
Fry WR, Dort JA, Smith RS, Sayers DV, Morabito DJ. Duplex scanning replaces arteriography and operative exploration in the diagnosis of potential cervical vascular injury. The American Journal of Surgery 1994; 168: 693-696.
Grant R. Vertebral artery concerns: premanipulative testing of the cervical spine. In: Grant R (Ed) Physical Therapy of the Cervical and Thoracic Spine. New York: Churchill Livingstone, 1994; 145-165.
Grant R, Johnson CL. The variability of measurement of vertebral artery blood flow. In: Proceedings of the Tenth Biennial Conference of the Manipulative Physiotherapists Association of Australia. Melbourne: Manipulative Physiotherapists Association of Australia, 1997; 60-62.
Grieve GP. Contra-indications to spinal manipulation and allied treatments. Physiotherapy 1989; 75: 445-453.
Griewing B, Morgenstern C, Driesner F, Kallwellis G, Walker ML, Kessler C. Cerebrovascular disease assessed by colorflow and power Doppler ultrasonography. Comparison with digital subtraction angiography in internal carotid artery stenosis. Stroke 1996; 27: 95-100.
Haynes MJ. Doppler studies comparing the effects of cervical rotation and lateral flexion on vertebral artery blood flow. Journal of Manipulative and Physiological Therapeutics 1996; 19: 378-384.
Hurwitz EL, Aker PD, Adams AH, Meeker WC, Shekelle PG. Manipulation and mobilization of the cervical spine. A systematic review of the literature. Spine 1996; 21: 1746-1760.
Ivancic JJ, Bryce D, Bolton PS. Use of provocational tests by clinicians to predict vulnerability of patients to vertebrobasilar insufficiency. Chiropractic Journal of Australia 1993; 23: 59-63.
Karnik R, Stollberger C, Ammerer H, Perneczky G, Slany J, Brenner H. Validity of continuous-wave Doppler sonography of the vertebrobasilar system. Angiology 1987; 38: 556-561.
Kenna C, Murtagh J. Back Pain and Spinal Manipulation. Sydney: Butterworths, 1989; 56-69.
Kuether TA, Nesbit GM, Clark WM, Barnwell SL. Rotational vertebral artery occlusion: a mechanism of vertebrobasilar insufficiency. Neurosurgery 1997; 41: 427-432.
Kunnasmaa KTT, Thiel HW. Vertebral artery syndrome: a review of the literature. The Journal of Orthopaedic Medicine 1994; 16: 17-20.
Lee KP, Carlini WG, McCormick GF, Albers GW. Neurologic complications following chiropractic manipulation: a survey of California neurologists. Neurology 1995; 45: 1213-1215.
Lewit K. Clinical picture and diagnosis of vertebral artery insufficiency. Journal of Manual Medicine 1992; 6: 190-193.
Licht PB, Christensen HW, Hojgaard P, Hoilund-Carlsen PF. Triplex ultrasound of vertebral artery flow during cervical rotation. Journal of Manipulative and Physiological Therapeutics 1998; 21: 27-31.
Lo LY, Ford CS, McKinney WM, Toole JF. Asymptomatic bruit, carotid and vertebrobasilar transient ischemic attacks--a clinical and ultrasonic correlation. Stroke 1986; 17: 65-68.
McDicken WN. Diagnostic Ultrasonics. Principles and Use of Instruments. 3rd ed. Edinburgh: Churchill Livingstone, 1991; 231-247.
Mikkonen RH, Kreula JM, Virkkunen PJ. Reproducibility of Doppler ultrasound measurements. Acta Radiologica 1996; 37: 545-550.
Nelson CF, Bronfort G, Evans R, Boline P, Goldsmith C, Anderson AV. The efficacy of spinal manipulation, amitriptyline and the combination of both therapies for the prophylaxis of migraine headache. Journal of Manipulative and Physiological Therapeutics 1998; 21: 511-519.
Nelson TR, Pretorius DH. The Doppler signal: where does it come from and what does it mean? American Journal of Radiology 1988; 151: 439-447.
Nilsson N, Christensen HW, Hartvigsen J. The effect of spinal manipulation in the treatment of cervicogenic headache. Journal of Manipulative and Physiological Therapeutics 1997; 20: 326-330.
Parker GB, Tupling H, Pryor DS. A controlled trial of cervical manipulation of migraine. Australian and New Zealand Journal of Medicine 1978; 8: 589-593.
Penning L. Review paper: differences in anatomy, motion, development and ageing of the upper and lower cervical disk segments. Clinical Biomechanics 1988; 3: 37-47.
Refshauge KM. Rotation: a valid premanipulative dizziness test? Does it predict safe manipulation? Journal of Manipulative and Physiological Therapeutics 1994; 17: 15-19.
Refshauge K. Testing adequacy of cerebral blood flow (vertebral artery testing). In: Refshauge K, Gass E (Eds) Musculoskeletal Physiotherapy. Clinical Science and Practice. Oxford: Butterworth-Heinemann, 1995; 125-131.
Rheault W, Albright B, Byers C, Franta M, Johnson A, Skowronek M, Dougherty J. Intertester reliability of the cervical range of motion device. The Journal of Orthopaedic and Sports Physical Therapy 1992; 15: 147-150.
Ringelstein EB, Zeumer H, Poeck K. Non-invasive diagnosis of intracranial lesions in the vertebrobasilar system. A comparison of Doppler sonographic and angiographic findings. Stroke 1985; 16: 848-855.
Rivett DA. Neurovascular compromise complicating cervical spine manipulation: what is the risk? The Journal of Manual and Manipulative Therapy 1995; 3: 144-151.
Rivett DA, Milburn P. Complications arising from spinal manipulative therapy in New Zealand. Physiotherapy 1997; 83: 626-632.
Rivett DA, Milburn PD, Chapple C. Negative pre-manipulative vertebral artery testing despite complete occlusion: a case of false negativity? Manual Therapy 1998; 3; 102-107.
Rivett DA, Reid D. Risk of stroke for cervical spine manipulation in New Zealand. New Zealand Journal of Physiotherapy 1998; 26: 14-17.
Rivett DA, Sharples KJ, Milburn PD. Effect of pre-manipulative tests on vertebral artery and internal carotid artery blood flow: a pilot study. Journal of Manipulative and Physiological Therapeutics 1999; 22: 368-375.
Shekelle PG, Coulter I. Cervical spine manipulation: summary report of a systematic review of the literature and a multidisciplinary expert panel. Journal of Spinal Disorders 1997; 10: 223-228.
Sinel M, Smith D. Thalamic infarction secondary to cervical manipulation. Archives of Physical Medicine and Rehabilitation 1993; 74: 543-546.
Steinke W, Meairs S, Ries S, Hennerici M. Sonographic assessment of carotid artery stenosis. Comparison of power Doppler imaging and color Doppler flow imaging. Stroke 1996; 27: 91-94.
Steinke W, Ries S, Artemis N, Schwartz A, Hennerici M. Power Doppler imaging of carotid artery stenosis. Comparison with color Doppler flow imaging and angiography. Stroke 1997; 28: 1981-1987.
Stevens A. Functional Doppler sonography of the vertebral artery and some considerations about manual techniques. Journal of Manual Medicine 1991; 6: 102-105.
Terrett AGJ. Vascular accidents from cervical spine manipulation: the mechanisms. Journal of the Australian Chiropractors' Association 1987; 17: 131-144.
Thiel H, Wallace K, Donat J, Yong-Hing K. Effect of various head and neck positions on vertebral artery blood flow. Clinical Biomechanics 1994; 9: 105-110.
Weingart JR, Bischoff HP. Doppler sonography of the vertebral artery with regard to head positions appropriate to manual medicine. Journal of Manual Medicine 1992; 30: 62-65.
Youdas JW, Carey JR, Garrett TR. Reliability of measurements of cervical spine range of motion--comparison of three methods. Physical Therapy 1991; 71: 98-106.
Young GR, Humphrey PRD, Shaw MDM, Nixon TE, Smith ETS. Comparison of magnetic resonance angiography, duplex ultrasound, and digital subtraction angiography in assessment of extracranial internal carotid artery stenosis. Journal of Neurology, Neurosurgery, and Psychiatry 1994 ; 57: 1466-1478.
Darren A Rivett, PhD
Associate Professor, Discipline of Physiotherapy, School of Health Sciences, The University of Newcastle, Australia.
Katrina J Sharples, PhD
Senior Lecturer, Department of Preventive and Social Medicine, University of Otago
Peter D Milburn, PhD
Associate Professor, School of Physiotherapy, University of Otago
ADDRESS FOR CORRESPONDENCE
Associate Professor Darren A. Rivett, Discipline of Physiotherapy, School of Health Sciences, The University of Newcastle, University Drive, Callaghan, NSW 2308, Australia. Telephone: +61 2 4921 7821 Fax: +61 2 4921 7902 Email: Darren.Rivett@newcastle.edu.au
Funding for this study was provided by the School of Physiotherapy, University of Otago.
Table 1. Comparison of measurements of physical characteristics of subjects. Data is presented as mean (SD) values, mean differences, SD of differences, and p-values. Physical characteristic First measurement Second measurement Systolic BP (mmHg) 132.5 (20.4) 123.7 (17.7) Diastolic BP (mmHg) 79.5 (13.8) 81.1 (10.1) Pulse rate (per min.) 68.0 (17.1) 65.3 (14.0) Extension range ([degrees]) 76.7 (16.1) 78.3 (16.3) Rotation range ([degrees]) 89.4 (6.7) 88.8 (5.6) Physical characteristic Difference (SD) p Systolic BP (mmHg) 8.8 (10.7) 0.002 Diastolic BP (mmHg) -1.6 (9.1) 0.44 Pulse rate (per min.) 2.7 (13.7) 0.39 Extension range ([degrees]) -1.6 (2.9) 0.02 Rotation range ([degrees]) 0.6 (3.6) 0.51 Table 2. Repeatability of ultrasonographic measurements taken at the atlanto-axial region of the vertebral artery in various neck positions as indicated by mean differences and 95% limits of agreement (LOA). Means (SD) of the averages of the scores are provided for interpretative purposes. Intraclass correlation coefficients (ICC) (2, 1) with 95% confidence intervals (CI) are also provided. Measure Neutral Peak systolic velocity (cm/s) Mean difference (LOA) -2.98 (-17.6, 11.7) Mean score (SD) 37.55 (8.94) ICC (95% CI) 0.69 (0.36, 0.87) End diastolic velocity (cm/s) Mean difference (LOA) -2.05 (-6.7, 2.6) Mean score (SD) 12.19 (2.95) ICC (95% CI) 0.61 (0.06, 0.85) Time averaged velocity (cm/s) Mean difference (LOA) -2.47 (-9.3, 4.3) Mean score (SD) 13.13 (2.93) ICC (95% CI) 0.40 (-0.03, 0.71) Lumen diameter (mm) Mean difference (LOA) -0.13 (-1.0, 0.8) Mean score (SD) 2.86 (0.50) ICC (95% CI) 0.66 (0.32, 0.85) Flow rate (ml/m) Mean difference (LOA) -11.67 (-47.5, 24.2) Mean score (SD) 54.42 (27.81) ICC (95% CI) 0.76 (0.37, 0.91) Resistance index Mean difference (LOA) 0.03 (-0.07, 0.13) Mean score (SD) 0.67 (0.07) ICC (95% CI) 0.72 (0.37, 0.88) Systolic/diastolic ratio Mean difference (LOA) 0.34 (-0.74, 1.42) Mean score (SD) 3.18 (0.70) ICC (95% CI) 0.68 (0.28, 0.87) Measure Extension Peak systolic velocity (cm/s) Mean difference (LOA) 2.05 (-15.6, 19.7) Mean score (SD) 42.55 (9.63) ICC (95% CI) 0.65 (0.31, 0.84) End diastolic velocity (cm/s) Mean difference (LOA) 0.76 (-8.4, 9.9) Mean score (SD) 14.44 (4.32) ICC (95% CI) 0.57 (0.19, 0.80) Time averaged velocity (cm/s) Mean difference (LOA) 0.20 (-10.7, 11.1) Mean score (SD) 14.95 (4.65) ICC (95% CI) 0.50 (0.07, 0.77) Lumen diameter (mm) Mean difference (LOA) -0.10 (-0.9, 0.7) Mean score (SD) 2.94 (0.55) ICC (95% CI) 0.79 (0.54, 0.91) Flow rate (ml/m) Mean difference (LOA) -4.15 (-54.3, 46.0) Mean score (SD) 64.20 (28.31) ICC (95% CI) 0.68 (0.35, 0.86) Resistance index Mean difference (LOA) 0.0 (-0.16, 0.16) Mean score (SD) 0.66 (0.08) ICC (95% CI) 0.58 (0.19, 0.81) Systolic/diastolic ratio Mean difference (LOA) 0.0 (-1.56, 1.56) Mean score (SD) 3.15 (0.81) ICC (95% CI) 0.63 (0.26, 0.84) Measure Rotation Peak systolic velocity (cm/s) Mean difference (LOA) 1.87 (-12.9, 16.6) Mean score (SD) 35.15 (11.74) ICC (95% CI) 0.82 (0.60, 0.92) End diastolic velocity (cm/s) Mean difference (LOA) -0.64 (-8.1, 6.8) Mean score (SD) 12.42 (3.99) ICC (95% CI) 0.65 (0.30, 0.84) Time averaged velocity (cm/s) Mean difference (LOA) -2.05 (-13.8, 9.7) Mean score (SD) 13.93 (5.27) ICC (95% CI) 0.51 (0.12, 0.77) Lumen diameter (mm) Mean difference (LOA) 0.04 (-0.8, 0.9) Mean score (SD) 2.90 (0.50) ICC (95% CI) 0.72 (0.42, 0.88) Flow rate (ml/m) Mean difference (LOA) -4.23 (-48.6, 40.1) Mean score (SD) 56.81 (26.24) ICC (95% CI) 0.70 (0.39, 0.87) Resistance index Mean difference (LOA) 0.03 (-0.11, 0.17) Mean score (SD) 0.66 (0.10) ICC (95% CI) 0.76 (0.47, 0.90) Systolic/diastolic ratio Mean difference (LOA) 0.37 (-0.73, 1.47) Mean score (SD) 2.89 (0.49) ICC (95% CI) 0.43 (0.00, 0.73) Table 3. Repeatability of ultrasonographic measurements taken at the C2/3 region of the vertebral artery in various neck positions as indicated by mean differences and 95% limits of agreement (LOA). Means (SD) of the averages of the scores are provided for interpretative purposes. Intraclass correlation coefficients (ICC) (2, 1) with 95% confidence intervals (CI) are also provided. Measure Neutral Peak systolic velocity (cm/s) Mean difference (LOA) -0.10 (-15.4, 15.2) Mean score (SD) 41.14 (12.65) ICC (95% CI) 0.84 (0.63, 0.93) End diastolic velocity (cm/s) Mean difference (LOA) 0.45 (-4.9, 5.8) Mean score (SD) 12.21 (3.28) ICC (95% CI) 0.72 (0.43, 0.88) Time averaged velocity (cm/s) Mean difference (LOA) 0.45 (-7.0, 7.9) Mean score (SD) 12.78 (4.22) ICC (95% CI) 0.68 (0.35, 0.86) Lumen diameter (mm) Mean difference (LOA) 0.10 (-0.5, 0.7) Mean score (SD) 2.99 (0.48) ICC (95% CI) 0.80 (0.56, 0.91) Flow rate (ml/m) Mean difference (LOA) 4.09 (-40.2, 48.4) Mean score (SD) 55.94 (23.95) ICC (95% CI) 0.65 (0.31, 0.85) Resistance index Mean difference (LOA) -0.01 (-0.13, 0.11) Mean score (SD) 0.70 (0.06) ICC (95% CI) 0.58 (0.20, 0.81) Systolic/diastolic ratio Mean difference (LOA) -0.06 (-1.56, 1.44) Mean score (SD) 3.47 (0.86) ICC (95% CI) 0.69 (0.36, 0.87) Measure Extension Peak systolic velocity (cm/s) Mean difference (LOA) 0.16 (-24.7, 25.0) Mean score (SD) 43.27 (9.42) ICC (95% CI) 0.41 (-0.05, 0.72) End diastolic velocity (cm/s) Mean difference (LOA) 0.28 (-8.2, 8.7) Mean score (SD) 14.25 (2.62) ICC (95% CI) 0.22 (-0.26, 0.60) Time averaged velocity (cm/s) Mean difference (LOA) 1.65 (-12.0, 15.3) Mean score (SD) 13.33 (3.08) ICC (95% CI) -0.10 (-0.51, 0.35) Lumen diameter (mm) Mean difference (LOA) -0.04 (-0.6, 0.5) Mean score (SD) 3.00 (0.47) ICC (95% CI) 0.84 (0.64, 0.93) Flow rate (ml/m) Mean difference (LOA) 6.30 (-67.8, 80.4) Mean score (SD) 58.90 (23.31) ICC (95% CI) 0.23 (-0.24, 0.60) Resistance index Mean difference (LOA) 0 (-0.16, 0.16) Mean score (SD) 0.66 (0.08) ICC (95% CI) 0.62 (0.25, 0.83) Systolic/diastolic ratio Mean difference (LOA) -0.01 (-1.51, 1.49) Mean score (SD) 3.12 (0.70) ICC (95% CI) 0.57 (0.17, 0.80) Measure Rotation Peak systolic velocity (cm/s) Mean difference (LOA) 0.72 (-16.5, 17.9) Mean score (SD) 35.61 (11.54) ICC (95% CI) 0.76 (0.49, 0.90) End diastolic velocity (cm/s) Mean difference (LOA) -0.16 (-7.8, 7.5) Mean score (SD) 12.34 (3.77) ICC (95% CI) 0.60 (0.22, 0.82) Time averaged velocity (cm/s) Mean difference (LOA) -0.25 (-10.1, 9.6) Mean score (SD) 12.38 (4.13) ICC (95% CI) 0.48 (0.05, 0.76) Lumen diameter (mm) Mean difference (LOA) -0.04 (-0.4, 0.3) Mean score (SD) 2.92 (0.51) ICC (95% CI) 0.94 (0.86, 0.98) Flow rate (ml/m) Mean difference (LOA) -0.83 (-48.0, 46.3) Mean score (SD) 51.78 (25.23) ICC (95% CI) 0.65 (0.30, 0.85) Resistance index Mean difference (LOA) 0 (-0.16, 0.16) Mean score (SD) 0.66 (0.10) ICC (95% CI) 0.73 (0.43, 0.89) Systolic/diastolic ratio Mean difference (LOA) 0.16 (-1.38, 1.70) Mean score (SD) 2.96 (0.70) ICC (95% CI) 0.54 (0.13, 0.79)
|Printer friendly Cite/link Email Feedback|
|Author:||Rivett, Darren A.; Sharples, Katrina J.; Milburn, Peter D.|
|Publication:||New Zealand Journal of Physiotherapy|
|Date:||Nov 1, 2003|
|Previous Article:||Tempting curiosity.|
|Next Article:||The effect of a combined exercise programme for people with Multiple Sclerosis: a case series.|