Traditionally a series of movements or positions thought to test the integrity of the vertebrobasilar arteries have been advocated prior to manipulation or mobilisation of the cervical spine (Maitland 1986; Grant 1994a; McKenzie 1990). Such cervical procedures have sometimes been associated with complications, very rarely of a serious nature, such as death or cerebrovascular accident. The aim of the test movements and certain direct questions is to try to identify patients for whom this type of treatment may be contraindicated. The topic is a controversial one. Some authorities are of the opinion that the risks of manipulation outweigh the benefits (Di Fabio 1999; Refshauge et al. 2002), and many consider the screening procedures unreliable and invalid (Dunne 2001; Rivett 2001; Gross and Kay 2001).
It has also been argued that the proof of a link between cervical manipulation and stroke is missing, as multiple case studies do not prove causation and a valid study to determine the true risk of manipulation would be impossible to conduct, as it would require millions of subjects (Chestnut 2004). In addition, several studies have postulated the incidence of stroke after neck manipulation; however, these estimates are hypothetical due to the lack of epidemiological evidence on the incidence rate of stroke in a representative population (Cote ei al. 1996) . The risk of neurovascular complication arising from cervical manipulation has been compared to the risk of serious gastrointestinal complication from NSAIDs for osteoarthritis (Dabbs and Lauretti 1995). The latter is more risky: 0.4% compared to 0.001 %.
Di Fabio (1999) reviewed 177 reports of injuries associated with cervical manipulation published between 1925 and 1997. The most common were arterial dissection, injury to the brain stem, cerebellar or spinal cord and Wallenberg syndrome. Death occurred in 18%. The majority of incidents were attributed to chiropractors; 41% of patients had received at least one other manipulation before the incident; rotation manipulations were the most common type of intervention; and the mean age of patients was 40.
Terrett (1998) reviewed a similar number of incidents with similar findings, but suggested that the younger age group and attribution to chiropractors simply reflected those most commonly attending for treatment and those most commonly providing manipulative treatment. Most disturbingly, some clinicians continued with further manipulation after the advent of symptoms associated with vertebroar-tery insufficiency (VBI). Ernst (2004) summarised the literature from 1995 to 2003 - in total more than three hundred patients had been reported, most commonly suffering stroke due to arterial dissection after cervical spinal manipulation.
Surveys of neurologists conducted in New Zealand and the UK demonstrate that complications following cervical manipulation, including strokes, have occurred without documentation (Rivett and Milburn 1997; Stevinson et al. 2001). The documented evidence thus probably underestimates the true extent of complications following manipulation (Terrett 1998), and complications appear as likely in the hands of physiotherapists as chiropractors (Rivett and Reid 1998).
The mechanism of injury is generally believed to be trauma to the vertebral artery around the level of the atlantoaxial joint (Grant 1994a; Terrett 1998; Mann and Refshauge 2001; Rivett 2004) . This section of the artery can be subject to excessive tension with the large range of rotation available at the C1 - C2 level and where the vertebral arteries are relatively fixed at the transverse foramina (Grant 1994b; Terrett 1998; Rivett 2004). A number of cadaveric studies demonstrated that certain movements caused a narrowing of the vertebral artery: cervical rotation in particular, extension with rotation, although not always extension only, and additional traction (Grant 1994b). It was this clinical and anatomical background that led to the establishment of test procedures to try to identify patients unsuitable for cervical manipulation. The vertebral arteries feed into the circle of Willis, providing less than 20% of cerebral blood supply, whereas the carotid arteries provide more than 80% (Grant 2002; Kerry 2005; Rivett 2004). This latter can also be affected by movement, especially extension (Rivett et al. 1999; Kerry 2005) (see carotid artery pathology section later).
Wallenberg's syndrome (Shelokov 1991) has been reported occurring as a result of a severely diminished flow in one vertebral artery; the decrease in flow can lead to the occlusion of the posterior inferior cerebellar artery on that side, resulting in a lateral medullary infraction. The most prominent clinical features are:
• dysphagia and ipsilateral palatal weakness (involvement of the nucleus ambiguous)
• impairment of sensation to pain and temperature on the same side of the face (involvement of descending root of the fifth cranial nerve)
Horner's syndrome in the ipsilateral eye (involvement of the descending sympathetic fibres)
• nystagmus (involvement of the vestibular nuclei)
• cerebellar dysfunction in the ipsilateral arm and leg (involvement of the restiform body and cerebellum)
• impairment of sensation to pain and temperature over the opposite half of the body (involvement of the spinothalamic tract).
Attempting to identify potential problems with vertebrobasilar insufficiency involves several components:
• physical examination tests (Table 8.8)
• awareness during treatment
° awareness following treatment.
Given the potential risky nature of the tests themselves, it is obviously better to try to identify at-risk patients before any examination or intervention is undertaken. Awareness of possible clues in the patient's history is therefore critical to safe management. If manual therapy is going to be performed, it is imperative to monitor the patient's response both during and after procedures, even when tests have been performed uneventfully
Certain signs and symptoms have been associated with vertebrobasilar insufficiency (VBI); these are listed in Table 8.5. It is important to remember that none of these is diagnostic of the condition, their diagnostic accuracy has not been tested, and VBI may not be the only cause of such a symptom. A useful tool to enhance memory of these signs and symptoms are the five Ds (dizziness, drop attacks, diplopia, dysarthria, dysphagia) and the three Ns (nausea, numbness, nystagmus).
Table 8.5 Clinical features associated with vertebrobasilar insufficiency or vertebral artery dissection
History • pain in head or neck
• sudden head/neck pain that has not been experienced before pain severe and sharp time delay between symptoms and features of brainstem ischaemia can be up to fourteen days.
dizziness/vertigo - most common nausea/vomiting facial paraesthesia - less commonly can involve trunk and limbs unsteadiness of gait/uncoordination diplopia extremity weakness - uncommon. hearing loss dysarthria dysphagia blackouts/fainting/drop attacks blurred vision/transient hemianopia tinnitus pallor and sweating. Source: Grant 2002; Furman and Whitney 2000; Terrell 1998; Thiel amd Rix 2005
Dizziness is a symptom with multiple causes. It is a common symptom in older populations, reported by 30% of people aged over 65 years (Colledge et al. 1996). Dizziness maybe caused by benign paroxysmal positional vertigo (BPPV), postural hypotension, a vestibular condition, labyrinthine concussion, a perilymphatic fistula, a mechanical cervi-cogenic condition as well as VBI (Furman and Whitney 2000; Wrisley et al. 2000). Possible causes of dizziness (Aspinall 1989):
• central (eg, demyelinating disease, tumour of the eighth cranial nerve, VBI)
peripheral (eg, benign paroxismal positional vertigo, vestibulopathy, meniere, cervical reflex vertigo)
systemic (eg, drugs/alcohol, hypotension, endocrine disease).
In the older population, the most common causes of dizziness are central vascular disease and cervical spondylosis, with postural
Clinical features -most common, in order of frequency
Other signs and symptoms hypotension andbenign paroxysmal positional vertigo being relatively unusual (Colledge et al. 1996).
Wrisley ei al. (2000) present clues for the different causes and a clinical reasoning algorithm. Cervicogenic dizziness is a diagnosis of exclusion, based on the exclusion of competing diagnoses, with the development of a robust test to demonstrate the cervical origin of dizziness being elusive (Wrisley et al. 2000). There is a test in which the head is stabilised and the body rotated, theoretically stimulating the neck proprioceptors and not the inner ear structures. However, this test has demonstrated poor specificity and sensitivity (Wrisley ei al. 2000). Consequently, to establish a relationship between dizziness and a cervical problem, the following points are recommended (Wrisley ei al. 2000):
• close temporal relationship between neck pain and dizziness both regarding onset and severity
• previous neck problems, possibly also with accompanying dizziness
• elimination of other causes of dizziness.
Again, many clues are to be found during the patient's history-taking.
Table 8.6 Differentiation between dizziness of cervical or other origin
Possibly cervical in origin Non-cervical in origin
Table 8.6 Differentiation between dizziness of cervical or other origin
Possibly cervical in origin Non-cervical in origin
Feelings of being pushed to one
Neck pain associated with dizziness
Limited cervical movement
Upper motor neurone signs and
Headache/upper limb symptoms
Source: Wrisley et til. 2000
Source: Wrisley et til. 2000
If the dizziness is associated with the neck pain in terms of onset, frequency and severity, and there are no other related features (Table 8.6), cervical origin is possible. Transient dizziness or spinning associated with changes in head position without neck movement, such as sitting up or turning over in bed, especially in the morning, may be due to benign paroxysmal vertigo. If a person answers yes to the following two questions, the Dix-Hallpike test should be used to rule out benign paroxysmal positional vertigo (BPPV) (Furman and Whitney 2000):
• Do you always have dizziness when you rise from lying to sitting?
Do you always have dizziness when rolling over in bed?
The Dix-Hallpike manoeuvre can be used to exclude dizziness from BPPV; however, for this the patient needs an adequate range of cervical movement. The patient is positioned in long sitting and the clinician rotates the head to 45 degrees and then brings the patient into supine quickly as they extend the head 30 degrees. If the patient cannot tolerate this manoeuvre because of pain, an alternative method is to have the patient in side-lying with the head rotated so their nose is pointing up and the back of the head is on the surface of the treatment table. The extension component is gained by lowering the end of the treatment table so that the patient's head falls into extension. Symptoms of spinning or signs of nystagmus indicate a positive test for benign paroxysmal positional vertigo (Wrisley ei al. 2000; Colledge et al. 1996; Lempert et al. 1995). Another test that has not been formally tested, but is said to indicate BPPV if it provokes severe dizziness, uses rapid head movements in different planes.
Various test protocols have been described, all with the aim of detecting patients who may have symptoms related to VBl (Aspinall 1989; Cote et al. 1996; Terrett 1998; Carey 1995; Barker ei al. 2000; APA 1988; Grant 2002; Magarey et al. 2004). Although there are minor variations to these pre-manipulation clinical tests, essentially they use the same manoeuvres, with end-range positional tests in rotation, extension, a combination of rotation and extension, and sometimes a position that mimics the manipulation position. The length of time that positions are sustained varies in different protocols, but in line with Australian Physiotherapy Association (APA 1988) test protocol, ten seconds is commonly given as the time. This is less if symptoms are evoked, and a gap of ten seconds should be included following each movement to allow for any latent response.
Concerns about the pre-manipulative cervical tests have been raised on several issues. By their very nature the tests are provocative. They attempt to provoke the symptoms one wishes to avoid, and therefore obviously may be dangerous in themselves (Di Fabio 1999); neurological complications due to testing have been reported (Rivett 2004). There have been reports of stroke induced by merely placing the head into the rotated position (Terrett 1998). The APA (1988) test protocol is time-consuming and not strictly adhered to even by manipulative therapists (Magarey et al. 2004). Perhaps most fundamentally is the research underpinning the test protocols, which is contradictory and ambiguous, whilst the reliability and validity of the test protocol remains uncertain (Assendelft et al. 1996; Kunnasmaa and Thiel 1994; Cote et al. 1996; Thiel and Rix 2005).
First, although it is likely that VBl test positions alter the flow parameters of the vertebral artery in some individuals, from a number of studies the evidence is contradictory and blood flow reduction does not appear to be a universal phenomenon (Magarey et al. 2004; Zaina et al. 2003; Rivett 2004; Thiel and Rix 2005). Whilst studies that use ultrasonography, Doppler ultrasound with real-time imaging or angiography have demonstrated reduced flow in some volunteers, almost an equal number of studies have found no differences in test positions (Rivett 2004; Magarey et al. 2004).
The link between flow parameters and symptoms has not been established. When significant reductions in blood flow have been demonstrated on contralateral rotation using Doppler sonography, no warning symptoms were elicited (Mitchell et al. 2004; Thiel and Rix 2005). Furthermore, case reports have identified individuals with false-negative symptom response to tests in the presence of occluded arteries (Bolton et al. 1989; Westaway et al. 2003). In a review of 321 studies in which blood flow reduction was matched with provoked symptoms, only 35 out of 274 'positive' tests induced symptoms, and conversely only 11 out of 47 'negative' tests were asymptomatic. Thus, the sensitivity and specificity of the tests in terms of correlation between blood flow response and symptom response were 13% and 23% respectively, making the test mathematically and clinically useless (Kerry 2005).
Cases have been reported of patients suffering serious complications after negative test results and after previously uneventful manipulations
(Terrett 1998; Rivett and Reid 1998). Apart from cadaveric studies, there is no evidence to suggest that if these tests were positive this indicates an underlying predisposition for VBI if a manipulation was performed (Terrettt 1998). Groups of patients reporting positive pre-manipulative tests have shown no decrease in blood flow in the vertebral arteries using Doppler ultrasonography (Licht et al. 2000, 2002; Thiel et al. 1994; Cote et al. 1996).
In a review of sixty-four patients who suffered cerebrovascular ischaemia shortly following cervical manipulation, no features in the history or examination, including the screening tests, allowed identification of an at-risk profile (Haldeman et al. 2002). Furthermore, a spontaneous onset vertebrobasilar artery dissection is more common than onset following manipulation (Haldeman et al. 1999).
The aim of the test protocol is to detect patients at risk of VBI prior to cervical manipulation or end-range mobilisation. The rationale is based on assumptions that (Rivett 2004):
1. positions of rotation and rotation/extension cause stenosis or occlusion of the contralateral vertebral artery
2. this causes reduction in blood flow through the vertebral artery
3. this will manifest itself in transient ischaemic signs and symptoms
4. the patient is unsuitable for vigorous manual therapy techniques as this might trigger VBl.
As the review has demonstrated, most of these assumptions are untenable (Thiel and Rix 2005). Stenosis or occlusion and disturbance of blood flow are not universal in the test positions; when blood flow is decreased this is rarely associated with symptoms, and people who have had previous manipulation or a negative response to the test have had a neurological event.
Thus, it is not currently possible to predict an at-risk patient prior to manipulation; even a negative test or previous uneventful manipulation is no guarantee that the procedure will be perfectly safe. The screening procedure may only indicate the patient's likelihood of survival if manipulation injures the vertebral artery (Mann and Refshauge 2001). Vertebrobasilar complications can occur spontaneously or with trivial force, but are also associated with cervical manipulation. To minimise risk it is prudent to employ the minimum force required to achieve the therapeutic ends using a progression of forces, and remembering that manipulation is not consistently better than other forms of treatment, including mobilisation or exercise (Refshauge et al. 2002). At-risk patient groups, who are more likely to suffer vascular incidents, may be better identified by history items, such as raised blood pressure, family history, smokers, overweight and so on (Kerry 2005). If anyone has clinical features that suggest VBI or vertebral artery dissection, provocative testing should not be performed and the patient should be referred appropriately.
"It is apparent that the validity of pre-manipulative testing is at best questionable, and its clinical value is limited; the capacity of the VA to withstand thrusting forces is not tested, although it may test the adequacy of the collateral circulation to maintain hindbrain perfusion" (Rivett 2004, p. 269).
"Provocative testing is very unlikely to provide any useful information in assessing the probability of manipulation induced vertebral artery injury" (Thiel and Rix 2005, p. 157).
Surprisingly, a large proportion of therapists who used manipulative therapy regularly did not provide information about the risks of the procedures, nor did they formally gain consent on every occasion (Magarey et al. 2004). Some even expressed strong opinions against the use of gaining consent, as it would likely put patients off if they knew the dangers involved. Such an attitude is an abandonment of legal and ethical duties that clinicians owe to patients. Legal situations vary around the world, but the contemporary philosophy of health care is much more consensual and based on informed patient participants than in the past. Provision of information and gaining consent for a technique that the patient cannot control and that involves a degree of risk is not simply recommended, but in many countries is a legal necessity (Magarey et al. 2004). The tests, "although poorly validated, seem to carry an important weight in court cases involving cerebrovascular injury after cervical spine manipulation" (Cote et al. 1996, p. 163).
"Health professionals have both an ethical and legal obligation to provide information and gain consentfor techniques such as cervical manipulation" (Magarey et al. 2004, p. 103).
After the death of a patient in September 1996 from thrombosis of vertebral artery and cerebellar infarction following a cervical manipulation, the Canadian coroner's court jury made the following recommendations:
• practitioners should obtain written informed consent an information sheet outlining risk of stroke should be provided
• provocative testing has not been demonstrated to be of benefit and should not be performed.
Providing information and gaining consent is as much a duty of care as the treatment itself, and failure to do so is a breach of this duty for which, in the changing legal climate, clinicians may well be sued for negligence (Refshauge et al. 2002; Magarey et al. 2004) .
There are reports of vascular accidents associated with non-manipulation scenarios that involve cervical rotation or extension, such as turning the head whilst driving, rap dancing, wrestling, archery, star gazing, neck extension during overhead work, radiography or a bleeding nose (Terrett 1998). On the whole, normal daily activities cause variations in vertebral artery blood flow that do not provoke symptoms; however, these cases reiterate the need for clinicians to be aware of the symptoms associated with VBI (Table 8.5) at all times during repeated movement testing. If patients report symptoms that are suggestive of vertebrobasilar problems, movement testing should proceed with caution. However, it should also be remembered that dizziness of cervicogenic origin might occur during movement testing - this is likely to lessen or abate with further repeated movement testing. If improvement does not occur, force progression should not be undertaken. If movement testing repeatedly provokes symptoms associated with VBI that do not improve, then testing should be abandoned and the findings reported to the patient's physician.
The first clues to help identify an at-risk patient might be picked up during the history-taking. The McKenzie system uses a progression of forces that starts with mid- to end-range patient forces prior to progressing to end-range, before therapist overpressures or mobilisations are even considered. Clearly this has an in-built safety mechanism. If very rarely individuals are prone to vertebral artery damage with relatively trivial forces, progression of forces ensures that such a response is likely to be recognised before major damage has occurred. Equally, this system of sequential force progressions allows time to establish the safety of one movement or level of forcebefore progression to the next level.
Both overpressure and mobilisation techniques can generate high levels of force and are not completely free of risk, with a case of stroke reported after a vigorous rotatory mobilisation (Michaeli 1993). However, most of the alarming accidents reported earlier were all associated with cervical manipulation. Only the final stage in the progression of forces uses manipulative procedures. These are only recommended if previous procedures in the same treatment principle have decreased but not abolished symptoms. If overpressure or mobilisation techniques have abolished symptoms, then the focus should be on repetition or avoidance of provocative postures. Before manipulation is even considered, the patient must have passed through several sessions in which lesser forces have been used.
If finally it is thought that manipulation is the appropriate intervention, a full discussion should be held with the patient about benefits, risks and alternatives and their written informed consent gained. If the patient gives informed consent to the procedure, end-range sustained positions should precede the manipulation. Position the patient in the pre-manipulative test position, question them about response, and return to neutral before the actual manipulation is performed. Never perform more than one manipulation per session, and do not repeat unless there is clear evidence of improvement.
If manipulation is performed, it should be remembered that rotation manipulation is particularly associated with VBI-type symptoms, especially ifforce is directed at the upper cervical levels. Lateral flexion, which has not caused loss of Doppler sounds indicating reduced vertebral artery flow (Terrett 1998), does not appear to be associated with these symptoms.
Table 8.7 Mechanical diagnosis and therapy and safeguards with VBI
all clinicians need awareness of VBI-associated symptoms (Table 8.5)
undertake a thorough history with specific closed questions about appropriate symptoms if indicated
• always use the progression of forces - test safety of movement and degree of force before progressing to end-range, overpressures or mobilisation monitor symptom response at all times enquire about any new symptoms never progress forces if VEI-associated symptoms are provoked only progress forces if transient dizziness has improved (no longer provoked) with repeated movements only progress to manipulation if all previous level of forces has decreased, but not abolished symptoms ensure that patient is fully informed about benefits, risks and alternatives of manipulation give space for patient questions and ensure patient consent is obtained perform end-range sustained tests as outlined below position patient in pre-manipulative test position and check response use only one manipulation in a session - monitor response never instigate further manipulation if no benefit gained or any adverse reaction provoked if manipulation is to be repeated on a subsequent occasion, consent and testing must be performed each time.
This is to be performed prior to each manipulation. Test protocol has been described in both sitting and lying; it is described here in lying (McKenzie 1990), but equally can be performed in sitting. The intensity and location of any symptoms are recorded prior to the performance of the test. The patient lies prone on the treatment table, leaning on the elbows and resting the chin on the outstretched fingertips with head maximally protruded and extended. Encourage the patient to relax so that the passive overpressure from the fingers allows maximum extension. The patient maintains this position for up to ten seconds, during which time they are asked to report any adverse effects or alteration in symptoms. Any provocation of adverse effects should terminate testing and contraindicate manipulation. On return to the neutral position, the patient is asked, "As a result of adopting that position, doyoufeel any nausea, dizziness or other effects?"
1f the patient is unaffected by this position the procedure is repeated, but with the addition of a rotation component, first to one side and then the other. The patient extends as before and then, whilst in extension, rotates as far as possible to one side and maintains this position for up to ten seconds. Again symptoms are monitored during and after the procedure. If the patient is unaffected by adverse symptoms the procedure is repeated to the opposite side, with symptom response again being monitored during and after the procedure. Finally, prior to the manipulation itself, the patient's head is positioned where the thrust would be performed; again this is maintained for up to ten seconds.
In the event that the patient becomes nauseous, dizzy or feels unwell during any part of the test procedure or afterwards, let the patient rest in the neutral position for several minutes until symptoms abate. The test movement may be repeated, but if the patient consistently reports adverse symptomatology, manipulation should be abandoned and the response and its possible implications reported to the patient's physician.
Table 8.8 Physical examination screening tests for patients prior to manipulation sustained extension sustained left/right rotation sustained extension and rotation simulated manipulation position positions are sustained for ten seconds (less if symptoms are evoked)
ten seconds in neutral position before next sustained posture if during any sustained position anysymptoms from Table 8.5 are provoked, position is abandoned and patient is contraindicated for manipulation.
Was this article helpful?