Table 3 in the technical version (http://ccachiro.org/cpg) provides details of studies’ specific manipulation techniques, and Table 4 in the technical version (http://ccachiro.org/cpg), exercise.

Table 4.   Aspects of treatment modalities that should be considered
to tailor treatment in response to modality-modifying risk factors

Aspects of modality that may reduce risk from risk factors

Manipulation (HVLA)     Force     Direction     Amplitude     Velocity     Patient position     Frequency     Duration     Location of treatment Mobilization     Force     Direction     Amplitude     Velocity     Patient position     Frequency     Duration     Location of treatment Ultrasound     Intensity     Duration     Wave frequency     Frequency of care     Duty cycle

Exercise     Repetition     Intensity     Frequency     Duration     Weight     Exertion     Type Traction     Direction of force     Harness     Weight     Frequency     Duration     Manual vs mechanical     Intermittent vs continuous Electro-therapies     Intensity     Duration     Frequency of care     Type of current or wave

Figure 1

Treatment recommendation 1.   Based on all the evidence below, we recommend the 3 sequential steps in the decision algorithm (Figure 1) – diagnosis (or assessment leading to diagnosis), treatment, reassessment – to treat patients with acute pain, an acute exacerbation of a recurrent pain, or chronic pain.

Similarly, we recommend the 3 sequential steps to treat patients with idiopathic pain or pain with an identified cause. The selection and dosages of treatment modalities will differentiate best practices for each unique combination of pain condition and patient. The selection and dosage of treatment modalities should respect recommendation 2.

Treatment recommendation 2.   Based on all the evidence below, we also recommend manipulation, mobilization, ischemic pressure, clinic- and home-based exercise, traction, education, low-power laser, massage, transcutaneous electrical nerve stimulation (TENS), pillows, pulsed electromagnetic therapy, or ultrasound – for patients with acute or chronic pain, where the origin of pain is known or unknown, to improve pain and some ROM – in dosages and methods based on the practitioner’s experience and the patient’s specific situation, as there is insufficient published evidence to support or refute narrow generalizations about the use of these treatment modalities.

Treatment recommendation 3.   Based on all the evidence below, in the absence of objective findings with neck pain not due to whiplash (e.g., ROM, muscle hypertonicity), we do not recommend that treatment be initiated. If, after a complete examination, all findings except for pain are normal, we recommend discharge of the patient from chiropractic care and, possibly, referral based on the practitioner’s experience.

Treatment recommendation 4.   Based on all the evidence below, in addition to the details of the 3-step sequence in recommendation #1, if home exercise is prescribed, we recommend frequent monitoring of its quality and a reassessment of the quality and effect of the home exercise after 2 to 4 weeks.

Immediate benefit from manipulation

Manipulation immediately improves pain and some ROM {L-4}, [11, 39–42] and a single manipulation ipsilateral to the location of pain is immediately better than a contralateral manipulation {L-2b}. [41] See Section 4.2 for the clinical importance of treatments with an immediate effect.


Short- and medium-term benefit from manipulation

Multiple manipulations improve pain in the short {L-4} [43–46] and medium {L-4} [43, 46–48] term, and some ROM in the short {L-4} [43, 44, 46, 48] and medium {L-4} [12, 43, 45–48] term. Manipulations in and opposite the direction of restriction may achieve greater ROM benefit than manipulation only in the direction of restriction {L-2b}. [44] Also, thoracic manipulations do not enhance the benefits from cervical manipulations in the short term {L-2b}. [48]

Treatment recommendation 5.   Based on the short- and medium-term benefit from manipulation, we do not recommend crossed bilateral transverse pisiform or anterior thoracic manipulations to be added to a course of cervical manipulations to improve pain and some ROM, unless where required for non-cervical benefits.

Short- and medium-term benefit from manipulation with stretching

Multiple manipulations, each preceded or followed by stretching of neck muscles, improve pain in the short and medium term {L-4}; [20] stretching either causes the relief or merely adds benefit to manipulation {L-5}.^GDC Multiple manipulations with or without {L-2b} stretching improve ROM in the medium term {L-4}. [20] Additionally, multiple manipulations may be better than unsupervised stretching {L-4}. [49]


Medium- and long-term benefit from manipulation with 3-point traction

Multiple manipulations with 3-point traction improve pain in the medium and long term {L-4}, [25, 26] and are better than no treatment {L-2b} [26] {L-4}, [25] but should be preceded by a detailed disclosure of the risks of the 3-point traction {L-5}.^GDC


Immediate benefit from mobilization

A single treatment of short mobilization sessions immediately improves pain {L-4}, [22] and is better than no treatment for an immediate effect {L-2b} (see Section 4.2).


Immediate benefit from ischemic pressure

Ischemic pressure applied to myofascial trigger points (TrgP) for 90 s immediately improves pain and some ROM {L-4} [50] (see Section 4.2).


Medium- and long-term benefit from exercise

Ongoing exercise improves: pain in the medium and long term {L-4}; [10] pain in the medium term for patients with unidentified conditions {L-4}, [32, 51–53] with strength exercise being the most consistently beneficial {L-1b}; [51] and pain and some ROM in the medium term for patients with cervical strain, herniated disc, degenerated discs {L-4}, [54] osteoarthritis, or “outset” resulting from minor injuries {L-4}. [31]

The evidence [27, 31, 32] is ambiguous about the advantage of exercise over no treatment, but suggests exercise is better than a placebo of clinical contact {L-1b}. [10] It is possible that subjectively perceived benefits such as pain reduction continue well beyond the treatment period, whereas objectively measured benefits do not {L-5}.^GDC


Medium- and long-term benefit from intensive or light exercise

Ongoing intensive or light exercise equally {L-2b} improve {L-4} pain in the long term, and intensive exercise is better than light exercise for objective outcomes in the medium term {L-2b}. [55]


Medium- and long-term benefit from exercise with education

Ongoing exercise with education improves pain in the long term {L-4}, but may be less beneficial than more passive treatment that includes medications for patients with cervical spondylopathy, soft tissue rheumatism, humeral tendinitis, tension neck, cervical syndrome, or rhizopathy {L-5}. [56]

Also, for patients without defined conditions, exercise-education combinations improve worst pain in the medium term {L-4} and have significantly greater effect than education alone {L-2b}; [57] this suggests the critical ingredient in exercise-education combinations is the exercise {L-5}.^GDC


Medium- and long-term benefit from exercise with education and home exercise

Ongoing exercise with home exercise and education improves pain and some ROM in the medium [13, 33, 34] and long [13] term {L-4}, and is better than extensive multi-modal treatments without exercise {L-2b}. [34]


Medium- and long-term benefit from exercise with multimodal treatments

Ongoing exercise with extensive multi-modal combinations improves pain and some ROM for patients with unspecified conditions in the medium [35, 58, 59] and long [58] term {L-4}.

It also improves pain and some ROM in the medium term for patients with osteoarthritis, cervical spondylopathy, strain, or cervical disc disease {L-4}. [60] In an exercise combination, exercise is likely most effective at improving objective outcomes, whereas the other modalities address subjective outcomes {L-5}.^GDC

There is conflicting evidence whether exercise is better than general medical care for pain in the medium {L-5}^GDC and long {L-2b} [33] term. Finally, the addition of more exercise to an exercise combination improves satisfaction and maintained benefit {L-5},^GDC but is not as good as extending the scope of the multi-modal treatment {L-4}. [59]

The evidence [35] is ambiguous about exercise’s advantage over no treatment {L-5}.^GDC


Summary exercise benefit statement

Multiple multi-modal treatments are the most effective and exercise, especially intensive exercise, is a critical element {L-5}.^GDC Overall, in addition to pain and ROM outcomes, ongoing strength exercise increases strength, ongoing endurance exercise increases endurance, but the outcome of coordination exercise is ambiguous {L-5}.^GDC

In addition, where it was reported, [28, 31, 32, 58] exercise was not consistently better than placebo or no-treatment groups {L-5};^GDC the propensity for neck pain to resolve on its own may be a confounding factor (see Section 4.1).


Short-, medium- and long-term benefit from home exercise with or without education or ultrasound

Home exercise may improve pain and some ROM in the medium and long term {L-4}, [29, 61] but extensive tailoring of the home exercise to each patient is required {L-5}.^GDC Some evidence {L-2b} [29] suggests that home exercise may be no better than no treatment for pain or ROM {L-5}.^GDC

Home exercise with education or monitoring improves pain and some ROM in the medium term {L-4}. [16] Home exercise with ultrasound improves pain in the short and medium term {L-4}; ultrasound enhances the effect of home exercise alone {L-2b}, [62] but not home exercise with massage {L-5}.^GDC However, home exercise with ultrasound is no better than no treatment for pain {L-2b}. [28]

Treatment recommendation 6.   Based on the summary exercise benefit statement and the short-, medium- and long-term benefit from home exercise with or without education or ultrasound, we do not recommend generic home exercise designed to improve pain or ROM that is not tailored to the individual patient.

We recommend tailored home exercise treatment, as rigorous as the patient can tolerate, if a loss of ROM, strength or endurance is found. It can be as frequent as once daily, with its rigor adjusted progressively.

Short- and medium-term benefit from ultrasound

Multiple ultrasound treatments improve pain and some ROM in the short and medium term {L-4}. [45]


Short- and medium-term benefit from low-power laser treatments

Multiple low-power laser treatments improve pain in the short and medium term {L-4}, [21, 23] and pain and some ROM in the short term for cervical osteoarthritis patients {L-4}. [63] The laser beam shows better results than placebo {L-1b} [21] {L-2b}, [23] which suggests that it causes the improvement {L-5}.^GDC


Medium-term benefit from pillows

Use of a cervical pillow during sleep improves pain in the medium term {L-4}. [64]

Treatment recommendation 7.   Based on the medium-term benefit from pillows, in addition to the details of the 3-step sequence in recommendation #1, we recommend a cervical pillow as a secondary treatment that should be initiated only after at least one cycle of diagnosis (or assessment leading to diagnosis), treatment and reassessment–and if prescribed, the pillow should be used nightly.

Short- and medium-term benefit from pulsed electromagnetic field therapy

Multiple pulsed electromagnetic field treatments improve pain and some ROM in the short and medium term {L-4}. [17, 18, 24] The pulsed electromagnetic field quality of the various vectors used (collar, electrode-points) shows better results than placebo {L-1b} [24] {L-2b}, [17, 18] which suggests that it is causal {L-5}.^GDC

Treatment recommendation 8.   Based on the short- and medium-term benefit from pulsed electromagnetic field therapy, in addition to the details of the 3-step sequence in recommendation #1, we recommend pulsed electromagnetic field treatment as an adjunctive, secondary treatment that should be initiated only after at least one cycle of diagnosis (or assessment leading to diagnosis), treatment and reassessment.

Immediate, medium- and long-term benefit from multimodal treatments

Multi-modal treatments improve pain and some ROM in the medium and long term {L-4}; [13, 15, 19, 34, 65] there is no discernable difference between combinations as long as they include home exercise, education, traction, 1 other secondary modality, and either manipulation, mobilization or clinic-based exercise {L-2b}. [13]

A single multi-modal treatment improves pain or pain and some ROM immediately {L-4}. [50] For example, one treatment of a 20-min hot-pack, active ROM, interferential current, and myofascial release (role of ischemic pressure undefined) is the first choice of 6 options of multi-modal treatments using secondary modalities to improve immediately: pain intensity, pain pressure tolerance, and pressure point thresholds (PPT) {L-5}^GDC (see Section 4.2).

The 6 treatments, from most effective to least, are:

• 20-min hot-pack, active ROM treatment, interferential current, myofascial release;
• 20-min hot-pack, active ROM treatment, stretch-and-spray, TENS;
• 20-min hot-pack, active ROM treatment, stretch-and-spray;
• 20-min hot-pack, active ROM treatment, ischemic pressure, TENS;
• 20-min hot-pack, active ROM treatment, ischemic pressure;
• 20 min hot-pack, active ROM treatment.

No additional benefit from magnets in necklaces

It is nearly certain that magnetic necklaces are no better than non-magnetic necklaces to improve pain in the short term {L-1b}, although improvement in pain follows the wearing of both {L-4}. [66]

Treatment recommendation 9.   Based on no additional benefit from magnets in necklaces, we do not recommend permanent magnet necklaces to improve pain, specifically because the monetary and lifestyle costs of a magnetic necklace do not appear to be counter-balanced by a clinical benefit.

No benefit from education or relaxation alone

Education alone does not improve pain [51, 57] or ROM [57] in the medium term {L-4}. Relaxation treatments are equal {L-2b} to advice to be active, in not improving pain or ROM in the medium or long term {L-4}. [27]

Treatment recommendation 10.   Based on no benefit from education or relaxation alone, we do not recommend education or relaxation alone to improve pain or ROM.

No immediate benefit from head retraction-extension exercise combinations alone

Head retraction-extension exercise combinations do not immediately improve pain {L-4}. [30]

Treatment recommendation 11.   Based on no immediate benefit from head retraction-extension exercise combinations alone, we do not recommend head retraction-extension exercise combinations to improve pain.

No immediate benefit from occipital release treatments alone

Occipital-release treatments do not immediately improve pain {L-4}. [30]

Treatment recommendation 12.   Based on no immediate benefit from occipital release treatments alone, we do not recommend occipital release treatments to improve pain.

4.1.   Natural history of neck pain

Men and women may experience neck pain differently. More men have acute pain, whereas more women have chronic pain. [67] Men also gain more pain relief from chiropractic treatment. [68] We agree with at least five reports [69–73] that pain is consistently associated with functional loss (e.g., altered patterns of neck muscle activation, reduction in strength and endurance) {L-5}, which may be without outwardly obvious structural change such as degenerative disease {L-2c}. [74]

Acute neck pain in adults is generally regarded to be self-resolving. One report [75] on patients with acute and chronic back disorder suggested that 90% of neck or back pain cases had self-resolved at 6 weeks {L-5}, corroborating another report. [34] Also, a review [76] on patients with neck pain suggested that more than 50% of patients experience a decrease in pain 2 to 4 weeks after onset and 80% will be asymptomatic in 2 to 3 months {L-5}. Another study suggested that patients (defined as chronic) who spent 16 weeks on a waiting list for physiotherapy experienced improved pain, cortical control, and 6 of 22 elements on an examination and physiologic test index {L-4}. [33]

We caution that there can be a structural or functional detriment associated with some patients’ pain, and we agree with at least one author [69] that the resolution of the pain may not signal a complete resolution of the detriment {L-5}. Specifically for these patients, even if treatment does not result in a faster or greater improvement in pain, good treatment will also address non-pain problems, which left untreated, may have permanent sequelae {L-5}.^GDC This justifies treatments that are expected to show less or slower improvement than the expected natural history of the treated pain in these patients {L-5}.^GDC

Authors of a Saskatchewan study [77] suggested that complete resolution of pain or disability is difficult to achieve (treatment status undefined) {L-2c}, and other authors reported that half of patients treated with primary care did not experience complete resolution after 1 year [78] or 5 years [79] (specific treatment type or consistency over the period undefined) {L-2c}. For patients whose pain is not self-resolving, we agree with at least 3 reports [68, 80, 81] that there may be a treatment opportunity to halt the evolution of acute pain to a chronic condition {L-5}, perhaps specifically for patients predisposed to chronic pain for non-pathologic reasons (low self-rated health and high levels of psychologic stress) {L-5}. [80] Therefore, we concluded that good practice does not universally mean waiting to find out whose pain resolves, and whose does not, before proceeding with treatment {L-5}.

We therefore recommend:

Treatment recommendation 13.   We do not recommend treatments that are expected to show less or slower improvement than the expected natural history of the treated pain in a particular patient, unless: a) the treatment also addresses non-pain problems that, left untreated, may have permanent sequelae, or b) it is deemed that treatment will halt the evolution of acute pain to a chronic condition.

Treatment recommendation 14.   If maximum clinical progress has been reached without all clinical goals being met, we recommend continuing care only if the patient chooses support or maintenance care. If all clinical goals have been met, we recommend continuing care only if the patient chooses “wellness” care.

In the recommendation above; Support Care means the necessary care for patients who have reached maximum therapeutic benefit, and for whom periodic trials of therapeutic withdrawal have led to deterioration and failure to sustain previous therapeutic gains. This form of care is initiated when the clinical problem recurs. [82]

Maintenance Care means elective care given at regular intervals designed to maintain maximum health and promote optimal function. It may incorporate screening procedures designed to identify a risk of developing problems pertaining to the patient’s health status and giving advice about this. [82]

Wellness Care means elective care given at regular intervals designed to maintain maximum health and promote optimal function.

4.2. The role of focusing on immediate clinical outcomes

Seven treatment studies [11, 22, 39–42, 50] focused on improved pain or ROM immediately after a single treatment. These treatments do not reflect a typical treatment plan. However, it is our understanding that an immediate post-treatment reduction in pain or increase in ROM suggests that the particular patient’s pain or ROM responds to the treatment and that subsequent treatments will reap further short-, medium- or long-term benefits {L-5}.

In addition, we consider that a legitimate role for a single treatment that provides an immediate relief of pain, if administered without temporarily increasing pain, is to permit another intervention that would otherwise be too painful for the patient to bear {L-5}.

Treatment recommendation 15.   We recommend the planned one-time use of a treatment specifically and only to determine the utility of further treatments or to permit the immediate use of an otherwise painful intervention, both purposes therefore requiring an immediately-subsequent patient assessment. Thus, we do not recommend the planned one-time use of a treatment to merely achieve an immediate clinical effect.

4.3. Multi-sectoral care

A key to successful chiropractic treatment appears to be integrated care that accommodates clinical sectors that may fall outside of some chiropractors’ practice. For example, unimodal and multimodal treatments that incorporate behavior modification [16, 29, 83] or pharmacologic interventions [12, 19, 56, 62, 75, 83–92] have shown benefit.

It is our understanding that integrated care means the integration of all modalities of treatment into an optimal care plan for a particular patient, including those chiropractic modalities supported by the evidence that is the foundation of this CPG {L-5}.

Treatment recommendation 16.   We recommend a concerted effort to mesh chiropractic care into that of other health disciplines to maximize patients’ gains from their chiropractic treatments (recovery from pain, impairment and disability, reduced costs, increased patient safety, increased satisfaction among patients and health care payers).

5.   Results: managing the risk of adverse events

The caliber of most studies in Section 5 was Level 5 (subjective extrapolation or observation, frequently based on a case study) and only a very few being Level 3 or better. Also, in general, the statistical significance of adverse event data was not reported. The results in Section 5 are of Level 5 caliber unless otherwise noted. The treatment section (Section 4) of this CPG was based almost exclusively on evidence from the literature; recommendations did not extrapolate beyond what the evidence clearly supported (except for the algorithms). To accommodate the caliber of research in Section 5, we relied on a greater level of the GDC’s (unpublished) practice expertise and cited it accordingly ({L-5}GDC).

Chiropractic regards the objective and subjective balance of benefits and the risk of adverse events to be especially important. The direct benefits of the treatment recommendations are expected to include those reported in the studies cited in Section 4, and reduced sick days and increased strength, endurance, flexibility, QoL, and ease of activities of daily living (ADL) – see technical version at http://ccachiro.org/cpg for details. Counter-balancing these benefits is the risk of adverse events.

For clarity, adverse events can be considered to be of 3 types:

• adverse events (AE) not associated with a treatment modality, but that occur in the clinical setting   (non-Tx-AE);

• adverse events associated with a treatment modality, but not a known or observable risk factor   (unforeseen-Tx-AE);

• adverse events associated with a treatment modality and predicted by an observable risk factor   (foreseen-Tx-AE).

These delineations are not static; for example, an adverse event that is associated with a treatment modality, but not predicted by a risk factor, may progress to being so (predicted by a factor) once this has been identified.

We deem that it is a chiropractor’s responsibility to understand and address all three types of adverse events (non-Tx-AE, unforeseen-Tx-AE, foreseen-Tx-AE) {L-5}.

Determination of the appropriate clinical response to an adverse event associated with a treatment, or to the relevant risk factor, is complicated by the manual nature of many chiropractic therapies. The risk of harm following treatment depends directly on the skill of the chiropractor {L-5}.^GDC Jagbandhansingh [93] echoed this conclusion in a 1997 review of USA malpractice claims:

“Chiropractic care in itself may not pose a clinical risk. However, treatment in association with lack of good professional judgement [or] failure to properly assess a patient’s condition may” (p. 64).

Risk-management, recommendation 17:   To manage the risk of adverse events associated with a treatment modality, if a chiropractor is uncertain about the caliber of any aspect of his or her technique with a particular patient, we very strongly recommend discontinuance of care and referral to colleagues until this is addressed.

5.1.   Managing the risk of adverse events not associated with a treatment modality, but that occur in the clinical setting (non-Tx-AE)

Chiropractors receive patients from all sectors of the population and may encounter any of the health problems that afflict everyone. Indeed, chiropractic patients are likely disproportionately ill compared with the general population. Côté et al. [94] reported that Saskatchewan patients seeking chiropractic care frequently presented with a wide array of serious comorbidities {L-2c}, such as heart disease.

When undiagnosed, these health problems may precipitate a non-Tx-AE before, during or after a treatment. Reports have described patients seeking care for symptoms treatable with chiropractic, but who have undiagnosed ailments outside of the scope of practice.

For example: the symptoms of various pathologies, such as:

systemic lupus erythematosus, Bornholm disease, [95]

neoplasms [95, 96] (osteoblastoma), [97]

cervical fracture, [98]

or internal carotid artery dissection [99] (ICAD) can mimic mechanical neck pain;

the symptoms of myocardial infarct (MI) can masquerade as neck pain [100] (and vice versa); [101]

seizures or transient ischemic attacks (TIA) can mimic migraines; [102]

neck and jaw pain can result from Marfan syndrome without spinal involvement; [103]

undetected septic arthritis can be harbored in patients complaining of joint pain while taking glucocorticosteroid therapy; [104]

and pain coincident with trauma can be paralleled by undetected cervical fractures, [105]

or (later) acute respiratory distress syndrome (ARDS) leading to death. [106]

Comorbidities can also be missed when overt signs are present, as when surface lesions in acute scalp lymphangitis indicated an untreated infection masquerading as cervical pain in at least one case. [107]

Attempts to systematically address the identification of serious comorbidities in advance of care may include using a 15-item self-report tool developed by Côté et al. [94] It is our understanding that a set of 15 presenting symptoms (adapted from McMillin) [108] should raise suspicion that the presenting cervical pain is not of mechanical origin {L-5}:

1. Trunk or lower extremity neurologic symptoms, especially long-tract signs.

2. Bilateral upper extremity pain.

3. Remote symptoms with neck movements (lower extremity).

4. Signs of sphincter dysfunction, bowel or bladder dysfunction or incontinence.

5. Fever, unrelenting nocturnal pain, weight loss, chronic fatigue.

6. Recent infection or surgery.

7. Polyarthralgia.

8. Dysphagia.

9. Nuchal flexion or extension rigidity, especially in the absence of trauma.

10. Cranial neurologic deficit or central nervous system symptoms.

11. Cervical pain related to general exertion (i.e., after climbing stairs).

12. Symptoms unchanged or progressive, despite previous functional management.

13. Onset of cervical pain associated with direct head trauma, loss of consciousness.

14. Sudden onset of cervical pain without trauma or incident.

15. Neck or occipital pain with a sharp quality and severe intensity, or severe and persistent headache, which is sudden and unlike any previously experienced pain or headache (see Section 5.3.1).

When a relevant comorbidity or a non-Tx-AE is noted, chiropractors have a responsibility to act in the best interest of their patients by immediate consideration of the situation, and referral to the appropriate healthcare resource as necessary.

Risk-management, recommendation 18:   Before, during or after treatment, we recommend immediate, in-depth consideration of possible explanations, and reconsideration of treatment options or referral to the appropriate health services when an adverse event (not known to be associated with a treatment) is noted; i.e., when a patient demonstrates signs or symptoms of an undiagnosed condition, or signs or symptoms not known to be associated with a treatment.

5.2.   Managing the risk of adverse events associated with a treatment modality, but not a known or observable risk factor (unforeseen-Tx-AE)

Managing the risk of unforeseen-Tx-AEs involves reacting appropriately if an event is noted. By definition, an unforeseen-Tx-AE follows the administration of a particular treatment, but there are no known factors differentiating patients at greater risk from others.

Thirteen [9–21] of the treatment studies reported about unforeseen-Tx-AEs (Table 2). Reports ranged from none [10–13, 16, 18, 20, 21] to those considered minor and self-resolving among a small fraction of study subjects, [14, 15, 16] or minor. [9, 19] The worsening of symptoms with treatment was also reported, [9, 10, 14, 64] but it was not possible to determine if this was as a result of treatment failure or an unforeseen-Tx-AE.

Richardson et al. [109] reported that about half of 98 patients had “some discomfort” following a 4-week trial of TENS {L-4}. About one-fifth had “unpleasant sensations” at or away from the TENS site, and a smaller number had headaches, muscle aches, nausea, bad temper, or dizziness. A systematic review [110] {L-1a} concluded that in a minority of patients, high-frequency (HF) TENS could cause skin rash whereas low-frequency (LF) TENS could cause a burning sensation over the area of electrodes, and HF-TENS or LF “train TENS” could cause skin irritation {L-5}.

Regarding manipulation; benign and transient unforeseen-Tx-AEs, or none, have been reported in a recent systematic review, [111] a comprehensive study [112] of more than 250 subjects, a prospective survey of 1,058 patients, [113] a prospective survey of 465 patients, [114] and 2 extensive literature reviews. [115, 116] In their 2003 study, Hurwitz et al. [112] reported that:

“No serious complications from spinal manipulation or manual treatment have been reported... [in] any of the published clinical trials involving [spinal] manipulation or mobilization” (p. 21).

Hurwitz et al. [112] suggested that manipulation conferred a greater risk than mobilization for benign, transient unforeseen-Tx-AEs that usually occur within 24 hours of treatment {L-5}.

Indeed, about 5 of every 100 patients treated with manipulation will feel “very noticeable” discomforts, whereas about 20 will feel moderate discomfort, and about 15 will feel slight discomfort {L-4}. [117] This total of affected patients (40 of 100) approximates the 34% to 55% estimate by Cagnie et al. [114] In more than 80% of these cases, discomforts disappeared within 24 hrs, and in more than 85%, discomforts occurred on the day of the manipulation {L-4}. [117]

Rare, unforeseen-Tx-AEs include spinal cord compression, facet edema, disk herniation, [116] long thoracic nerve palsy, [118] ruptured cervical discs, [119] and diaphragm paralysis related to phrenic nerve trauma. [120] In addition, women and younger patients (27 to 46 years) have reported more complaints {L-2c}. [113]

Risk-management, recommendation 19:   During or after treatment, we recommend heightened vigilance for adverse events associated with a treatment modality, but not a known or observable risk factor (unforeseen-Tx-AE) when a relevant treatment is planned or administered–and immediate, in-depth consideration of possible explanations, and reconsideration of treatment options or referral to the appropriate health services when an event is noted.

Without considering the issue of dissection, patient differences in susceptibility to adverse events related to manipulation are apparent (being female, being older, smoking, regular medication use, history of migraine) {L-4}. [114] However, we agree with the conclusion suggested by the results of Cagnie et al. [114] that these risk factors have not been consistently predictive, and we question their clinical utility {L-5}.


5.3.   Managing the risk of adverse events associated with a treatment modality and predicted by an observable risk factor (foreseen-Tx-AE)

Managing the risk of foreseen-Tx-AEs involves respecting risk factors and reacting appropriately if an adverse event is noted.

We extracted a list of factors from the risk management literature we found (more than 6,774 citations, more than 309 relevant articles), which likely constituted all reported foreseen-Tx-AEs. This list reflected our understanding of each relevant study’s determination of whether, strictly within the context of the study, the factor was a contraindication to the studied treatment modality, or merely indicated a requirement for caution {L-5}.^GDC

We concluded that most of the contraindications we had listed were likely not contraindications outside of the specific study context or with patients other than the studied cases {L-5}.^GDC Many of the comorbidities Côté et al. [94] reported in their survey of 1,131 Saskatchewan patients seeking chiropractic care {L-2c} were considered contraindications in this list.

These included

arthritis   (26.8% of patients seeking a chiropractor only, 50% of patients seeking a chiropractor and medical physician),

cardiovascular disease   (12.2%, 25%)

and significant depressive symptoms   (22.5%, 38.5%).

Most of the ailments that mimic mechanical pain for which patients may seek chiropractic treatment (Section 5.1) were also considered contraindications in this list:

systemic lupus arythmetous,

Bornholm disease   [95] or

neoplasms   [95, 96]   (osteoblastoma),   [97]

seizures, TIAs,   [102]

Marfan syndrome,   [103]

undetected septic arthritis,   [104]

and cervical fractures.   [105 ]

With most patients presenting with one or more of the factors we listed as a contraindication or caution, we deem that best practice rests on differentiating these factors into:

absolute contraindications   (Section 5.3.1),

factors requiring modification of a desired treatment modality   (Section 5.3.2),

or factors suggesting caution   (Section 5.3.3).

Risk-management, recommendation 20:   We recommend respecting the absolute contraindications listed in Tables 3a to 3h, and the best-practice patterns of absolute contraindications, treatment modality modification and caution described in Sections 5.3.1, 5.3.2 and 5.3.3 of this CPG.

As mentioned in Section 3, individual modalities of treatment generally differed between studies (see Tables 1 and 2). In addition, treatments were often vaguely defined (e.g., “manual therapy”). These factors dramatically reduced our ability to synthesize results across studies, and undermined the usefulness of these studies for practitioners.

Research, recommendation 22:   We very strongly recommend the inclusion of exact, reproducible descriptions of treatments in chiropractic studies, to ensure these are relevant and useful to the practicing chiropractor.

6.2.   No-treatment comparison groups

Only 12 of the studies [22, 25–35] that supported the treatment recommendations (Section 4) included a no-treatment comparison group. The evidence that showed chiropractic treatment to be clearly better than no treatment was for very specific examples of manipulation with traction or mobilization, where the primary clinical goal was to reduce pain.

In light of the potential for self-resolution of pain complaints (Section 4.1), this weakness in a study is fatal to the utility of the study’s results in treating pain.

Research, recommendation 23:   Where ethical, we strongly recommend the inclusion of a no-treatment group in comparative studies of chiropractic treatments, to ensure these are useful to the front-line chiropractor.

6.3.   Placebo comparison groups

In research, a placebo is generally considered an intervention that is used in across-group comparisons, to separate the direct benefits of the studied treatment from the effects of personal contact, the clinical milieu and patients’ belief in these benefits – thereby isolating the critical element of the studied treatment. Only 710, [17, 18, 21–24] of the treatment studies (Section 4) included what their authors considered to be placebo groups. Of these 7 studies, we consider only 3 studies [21, 22, 24] to have used effective placebos. The others used placebos that included more than the factors defined above (palpation, [23] physical contact with treatment device, [17, 18] infrared irradiation [10]).

This merely reflects how difficult it is to design valid placebos in a chiropractic practice laden with physical contact modalities that are, by design, tailored to each patient’s unique needs. [122] In addition, research is susceptible to including placebos that are an incremental part of the studied treatment (e.g., palpation compared with palpation plus acupuncture) [123] or an independent treatment modality (e.g., ”manual contact” with no segmental movement compared with mobilization). [22]

Pain is subjective [124, 125] and strongly influenced by psychologic, [80, 124, 126] cultural, [126] physiologic, [80, 124, 127] family, and work [127] factors, and the patient’s belief in the effectiveness of a treatment. [128] Pain is likely very susceptible to personal contact, the clinical milieu and belief in the benefits of a treatment. Several studies bear this out:

• Pain intensity, pain frequency, disability, sick-leave taken, general health, distress, and risk improved after 8 weeks of placebo (twice-weekly, ”lowest level” ultrasound), and remained so at 6 and 12 months {L-4}. [58]

• After 3 to 5 weeks of placebo (eighteen 30-min de-activated pulsed electromagnetic field [PEMF] “MT System” treatments, 3 to 5 per week) osteoarthritic patients experienced improved: pain, limitation of rotation and ease of ADL midway through treatment, at the end of treatment and 1 month later; tenderness midway through treatment and at the end of treatment; and pain on passive motion at the end of treatment {L-4}. [18]

• Pain-related disability improved after 6 weeks of placebo (twice-weekly infrared irradiation), and remained so 6 months later {L-4}, equal to the same regimen with exercise at each treatment {L-1b}. [10]

• The beneficial effect of placebo is likely illustrated in a study [66] of 3 weeks of continuous wearing of a magnetic or a non-magnetic (placebo) necklace. Both groups experienced significant improvement {L-4} on subjective evaluations of pain intensity and frequency, with no difference between the groups {L-1b}.

The above, plus our understanding of the holistic approach of chiropractic and the nature of chiropractic modalities, has led us to conclude that what is generally termed placebo effect (effect of contact, milieu, belief) is a valid and integral part of chiropractic treatment of neck pain. This suggests that best practice includes considering contact, milieu and belief as an active treatment modality that can be refined. One repercussion is that future chiropractic research should consistently and specifically examine this.

Research, recommendation 24:   We very strongly recommend the inclusion of well-delineated treatment groups that isolate placebo effects (effect of personal contact, the clinical milieu and patients’ belief in the benefits of a treatment) in comparative studies of chiropractic treatments, to create a solid evidence base for future chiropractic.

7.   Implementing the recommendations  

The following information tools are presented to aid in implementing this CPG:

a summary Table of the evidence-based cervical pain benefits of chiropractic treatment (Appendix 3);

algorithms that illustrate the process of individualizing care (Figure 1) and managing the risk of dissection (Figure 2, discussed in Appendix 1);

and a clinical question and answer list (Section 7.1 below).

As well, this CPG is reinforced by the extensive dissemination, implementation, evaluation, and revision activities described in the development, dissemination, implementation, evaluation, and revision plan (DevDIER). [6]

8.   Conclusion  

Well executed research, and the high caliber evidence this can produce, can objectively quantify outcomes subjectively observed in daily practice. This research can be used to confidently predict the response to a treatment for a clearly delimited problem. Thus, evidence-based practice permits the administration of interventions that are objectively known to produce a favorable clinical response in patients sharing the same problem.

This CPG examined the evidence for treatments designed to alleviate neck pain. Pain is fundamentally qualified by its multi-factor origin and the idiosyncratic way it demonstrates itself as a set of symptoms. Thus, we agree with at least one literature review, [125] examining the use of research in chiropractic practice, that it is likely impossible for the subjects of any studied group to be experiencing the “same” pain condition – homogeneously susceptible to a tested treatment {L-5}.

One repercussion is that a study assessing the impact of a treatment’s effectiveness is unlikely to satisfactorily direct good practice on its own, because it is likely other patients’ pain (and thus treatment needs) will differ from the studied sample. The GDC concluded this is a unique confounder in using evidence-based chiropractic to treat pain.

Faced with this, we consider that the most effective method of using the scientific literature in practice includes the application of “individualizing” algorithms. These illustrate the optimal decision path used to tailor objectively-determined treatments, which benefit the average patient, to particular patients – the care that is ultimately offered is thereby unique to each patient.

Research recommendation 25:   We recommend, in future research about the chiropractic treatment of pain, a concerted effort to develop evidence-based algorithms that illustrate the optimal decision path used to individualize care; our recommended method to incorporate evidence about treatments that are objectively known to benefit the average patient.

1.1.   Incidence of dissection or VAD or VAD-related stroke associated with manipulation

The evidence [149, 156, 161, 164–167] suggests that VAD following manipulation occurs very rarely (1 per million manipulations [Appendices 16a, 16b in the technical version of this CPG available at http://ccachiro.org/cpg]), and that most VAD arise from ordinary ADL, but none of the evidence is of sufficient quality to support objective conclusions {L-5}.GDC The core difficulty appears to be that if the rate of manipulation-associated VAD is 1 per million manipulations, 12 million manipulations would have to be studied to reach a scientifically confident conclusion {L-5} [167] – a logistically impractical proposition at this time {L-5}.^GDC


1.2.   Role of manipulation

The evidence suggests that more manipulations do not mean more complications {L-4} [148] and manipulation does not exacerbate the outcome of dissections {L-4}. [149] However, the evidence also suggests there is a statistical association between manipulation and bilateral VAD [149, 153, 156, 168] (with VAD symptoms moments to 10 days later), but not unilateral VAD {L-4}, [149] and between manipulation and VAD within 30 days {L-4}. [153]

Contrasting one study {L-4} [149] but supported by another {L-4}, [168] we deem that it is unlikely (based on anatomic principles) that manipulation can be responsible for bilateral VAD but not unilateral {L-5}, and thus an association between manipulation and bilateral VAD is likely not clinically important {L-5}.^GDC

Also, as suggested by Hufnagel’s results {L-4}, [156] it is unlikely (based on physiologic principles) that a manipulation can be responsible for a VAD more than 3 weeks later {L-5};^GDC thus, a 30-day association between manipulation and VAD is likely not clinically important {L-5}.^GDC

Considering the possibility that manipulation may cause dissection directly, we strongly agree with Sackett’s [167] suggestion that none of the available evidence shows causation. However, the evidence does suggest that patients with impaired vertebral artery flow may seek manipulative care for symptoms of these impairments {L-4} [148] {L-5}. [149, 150, 154]

Research, recommendation 27.   In studies of the association between manipulation and dissection, we very strongly recommend discriminating between realistic and unrealistic post-manipulation periods to ensure findings are relevant to drawing conclusions.

1.3.   Managing the issue of dissection

Irrespective of cause, it appears there may be a remote risk of VAD occurring subsequent to manipulation {L-5}.GDC We agree with Magarey et al. [146] that this risk is a clinical factor to be managed, much as drug side effects are in pharmacotherapy and anesthetic adverse events are in surgery {L-5}. This risk does not suggest that manipulation should be excluded from the armamentarium of practice as several authors [169, 170] have hinted {L-5}.^GDC

The diagnosis (or assessment leading to diagnosis), treatment and reassessment steps of the decision algorithm in Figure 1 guided our description of the factors relevant to managing the risk of dissection:

informed consent   (Section 1.3.1 of Appendix 1);

pre-disposing history   (Section 1.3.2 of Appendix 1);

predisposing factors in physical assessment   (Section 1.3.3 of Appendix 1);

and the occurrence of dissection   (Section 1.3.4 of Appendix 1).

1.3.1.   Informed consent

The evidence suggests that informed consent is as important for younger and healthier patients as it is for others, because these patients bear a disproportionately high burden of dissection-related stroke {L-3b} [157] {L-4}. [153]

Risk-management, recommendation 28   We very strongly recommend obtaining informed consent based on current evidence, and respecting the 3 sequential steps in the decision algorithm (Figure 1) – diagnosis (or assessment leading to diagnosis), treatment, reassessment – when caring for any patient.

1.3.2.   Predispositions to dissection in a patient’s history

Trauma   The evidence suggests that a recent history of trauma may predispose to dissection {L-4}. [149]

Lifestyle   The evidence suggests that among lifestyle factors, smoking may predispose to VAD [149] among those who are vulnerable to tobacco’s negative, cumulative vascular effects [153] {L-4}. The factor governing this vulnerability is unknown {L-5}.^GDC

Sex   The evidence suggests that a person’s sex is not a determinant of non-traumatic VAD {L-4}. [168]

Non-vascular illnesses   The evidence suggests that a history of non-vascular illness is not a definitive determinant of VAD {L-4}. [153, 156]

Vascular pathologies (excluding dissection or impaired vertebral artery flow)   The evidence suggests that a history of vascular pathologies is not a definitive determinant of VAD {L-4}, [156] although it may be suggestive of risk; of the vascular pathologies that have been proposed as risks for dissection, the most clearly demonstrated have been tissue abnormalities of the cervical arteries {L-3b} [171] {L-4}, [172] but the link between these abnormalities and dissection, and the practicality of pre-manipulatively identifying these remains elusive {L-5}.^GDC Patients predisposed to dissection by these factors may seek chiropractic care for neck pain {L-5}.^GDC

Hyperhomocysteinemia and cervical artery dissection   The evidence suggests that hyperhomocysteinemia is associated with CAD - [173] or dissection - [174] related stroke {L-3b}, but not CAD per se {L-3b}, [173] contradicting at least one extensive review. [175] The role of 5, 10-methylenetetrahydrofolate reductase (MTHFR, a deficient enzyme form associated with hyperhomocysteinemia) is ambiguous {L-5};^GDC it is not associated with dissection-related stroke (a mix of VAD and CAD) {L-3b} [174] or non-dissection stroke {L-5}, [175] but it is associated with CAD-related stroke {L-3b}. [174] Hyperhomocysteinemia may be associated with VAD, but results are too inconclusive at this time for this to have clinical importance {L-5}.^GDC

Historical predispositions; caution or contraindication?   Together, the current evidence [144, 148–150, 153, 156, 158, 169, 171–179] suggests that none of the predisposing factors hypothesized in the literature definitively predict a dissection-related “cerebrovascular ischemic event” {L-4}, [144] and, therefore, none is a contraindication to manipulation {L-5}.^GDC However, other evidence suggests that caution should be exercised in treating patients with a history of trauma {L-4}, [149] a smoking habit {L-4} [149] (especially among patients vulnerable to tobacco’s negative, cumulative vascular effects {L-4}), [153] or known arterial tissue abnormalities {L-3b} [171] {L-4}. [172]

Risk-management, recommendation 29   We recommended caution in treating a patient with trauma, a smoking habit or known arterial tissue abnormalities to manage the risk for dissection, but the evidence does not warrant that these be contraindications to manipulation.

1.3.3.   Noting predispositions during physical examination; impaired vertebral artery flow

Doppler identification of impaired vertebral artery flow   The evidence of an extensive review of the literature suggests that a positive (impaired) Doppler “flow test” does not predict impaired vertebral artery blood flow {L-5}. [146]

Provocative identification of impaired vertebral artery flow   Pre-manipulative vertebral artery function tests to identify patients with impaired flow have been part of practice knowledge since Smith and Eldridge introduced a test in 1962. [147] Since then, several tests have been developed, and the vertebral artery flow effects [180] of these are moderately well understood: Barré-Leiou’s sign test, George’s cerebrovascular craniocervical functional test, Maigne’s test, Hautant’s test, Underberg’s test, Dix-Hallpike maneuver (also known as Nylen’s or Barany’s maneuver), and deKleyn’s test. [147] The most common is deKleyn’s test. [147]

The evidence suggests that a positive (impaired) provocative “flow test” rarely indicates changes in vertebral artery blood flow {L-4} [147] and, consequently, deKleyn’s test is neither sensitive nor specific. Thus a positive test should not be an absolute contraindication to manipulation {L-5}. [147] Other results {L-4} have also suggested pre-manipulative flow testing is unlikely to identify patients with flow impedances {L-5}. [146]

Observational identification of impaired vertebral artery flow (overt symptoms of VBI)   We agree with the evidence defining the signs and symptoms of VBI (nystagmus, nausea, numbness, diplopia, drop attacks, dysphagia, dysarthria, ataxia), [181] as differentiated from benign paroxysmal positional vertigo (BPPV) {L-5}. [146] It is our understanding that these differentiated signs and symptoms may indicate the vertebral artery has an anatomic predisposition to dissection, has a narrow safety margin for the impairment of flow, or is dissected {L-5}. This suggests caution in proceeding with treatment when overt signs or symptoms of VBI are noted {L-5}.^GDC

Risk-management, recommendation 30.   We recommend an assessment for signs and symptoms of unprovoked VBI (differentiated from BPPV) to identify the possibility of impaired vertebral artery flow (signs and symptoms are: nystagmus, nausea, numbness, diplopia, drop attacks, dysphagia, dysarthria, and ataxia), because we recommend caution in treating a patient with suspected impairment of flow. However, the evidence does not warrant this being a contraindication to manipulation.

Risk-management, recommendation 31.   We do not recommend an assessment for signs or symptoms of unprovoked VBI (differentiated from BPPV) to identify the presence of dissection, or to identify patients with greater or lessor risk of symptomatic (ischemia-provoking) dissection subsequent to manipulation; the assessment lacks predictive value.

Risk-management, recommendation 32.   We do not recommend Doppler or provocative pre-manipulative vertebral artery function tests (e.g., deKleyn’s test) to identify impaired vertebral artery flow, the presence of dissection, or patients with greater or lesser risk of symptomatic (ischemia-provoking) dissection subsequent to manipulation; the assessment lacks predictive value.

1.3.4.   Dissection in the chiropractic clinic

Identifying the occurrence of dissection before or during a visit   We agree with at least 3 other reports [153, 162, 182] that manipulation is contraindicated for patients with an active or existing VAD or CAD, although there have been uneventful case reports of manipulation being used with benefit in treating patients who have “recovered” from dissections. [169, 183]

Risk-management, recommendation 33.   We do not recommend manipulation for patients who present with active or existing VAD or CAD.

This is relevant because impaired vertebral artery flow can exhibit symptoms [149, 150, 154] that may lead to the patient seeking chiropractic care {L-5};^GDC VAD and CAD are associated with neck-pain and headache {L-4}, [99, 149] although asymptomatic ICAD has been reported {L-5}. [162]

The evidence suggests that symptoms of impaired vertebral artery flow can lead patients to seek chiropractic care {L-4}; [149, 156] and that in most patients with dissection-precipitated cerebral ischemia, the event of dissection [161] – or the presence of dissection leading to stroke [148, 153] (specifically VAD) [149, 168, 184] – is associated with neck or occipital pain with a sharp quality and severe intensity, or a severe and persistent headache, which is sudden and unlike any previously experienced pain or headache {L-4}. However, this pain may not be clearly distinguished from musculoskeletal or neuralgic pain {L-5}. [144, 153]

Risk-management, recommendation 34.   We recommend caution in treating a patient who reports a recent (but not ongoing) neck or occipital pain with a sharp quality and severe intensity, or a severe and persistent headache, which was sudden and unlike any previously experienced pain or headache (even when it is suspected the pain was of a musculoskeletal or neuralgic origin).

Risk-management, recommendation 35.   We recommend immediate discontinuance of treatment and referral to emergency health services when a patient complains in the course of care (diagnosis [or assessment leading to diagnosis], treatment, reassessment) of neck or occipital pain with a sharp quality and severe intensity, or a severe and persistent headache, which is sudden and unlike any previously experienced pain or headache (even when it is suspected the pain is of a musculoskeletal or neuralgic origin).

Mitigating the harm of VAD; a stroke   The full extent of identifying and managing the risk of stroke subsequent to VAD is beyond the scope of this CPG. However, we conclude that any sign of VAD should precipitate immediate referral to emergency health services {L-5}.^GDC

The evidence suggests that progression from VAD to stroke may be indicated by specific neurovascular signs or symptoms:

unilateral facial paresthesia,

objective cerebellar signs,

lateral medullary signs,

and visual field defects {L-4}. [168]

Vertigo is also cited, [168] but we consider this to be a prevalent symptom of various less alarming syndromes, such as those with inner ear involvement {L-5}.^GDC We agree with the text by Goetz [185] that provides a more complete list of signs and symptoms {L-5}.

Risk-management, recommendation 36.   We recommend immediate discontinuance of treatment and referral to emergency health services when, in the course of care (diagnosis [or assessment leading to diagnosis], treatment, reassessment), a patient demonstrates at least 1 of 4 signs or symptoms of neurovascular impairment (unilateral facial paresthesia, objective cerebellar signs, lateral medullary signs, visual field defects) or other signs or symptoms of neurovascular impairment with unknown cause, irrespective of complaints of neck or head pain. In addition, we recommend immediate investigation for these 4 signs or symptoms of neurovascular impairment whenever a patient demonstrates vertigo – if none are present, we recommend caution in treating the patient because of the continued risk for neurovascular impairment.

1.3.5   Recommendation summary

We recommend management of the risk of VAD subsequent to manipulation in keeping with the decision algorithm of Figure 2, to maximize the demonstrated benefits and minimize the theoretic harms associated with manipulation.

We recommend a risk-management approach that includes:

Appendix 2:   Stroke; an adverse event of the rotation component of manipulation?  

In this CPG, unless otherwise noted: dissection means cervical (carotid or vertebral) artery dissection. Where not noted, the results in Appendix 2 are of Level 5 caliber.

Stroke has been associated with chiropractic manipulation in the news media for some time. Reports associating chiropractic treatment with stroke rest on accepting a hypothetical 5-link chain:

1) manipulation may be associated with dissection or cervical artery intima injury with

2) possible clot formation at the site of injury,

3) from which emboli or thrombi may be shed,

4) which in-turn may occlude arteries feeding the brain,

5) possibly causing a full stroke. However, not all dissection leads to stroke, and not all stroke is the result of dissection.148,149,156–160

The evidence suggests that rotary manipulation is very popular but its role in precipitating cerebrovascular (CV) accidents remains unclear {L-5}.^GDC On one hand, no particular manipulation technique predicts stroke {L-4}, [148] and no particular technique is disproportionately prevalent among manipulation-associated stroke patients {L-4}. [156]

On the other hand, rotation techniques applied to the upper-cervical spine are more associated with CV accidents than non-rotation or rotation to the mid- or lower-cervical spine {L-4}, [186] and upper-cervical manipulations of any type are 4 times more associated with cerebrovascular incidents (CVIs, transitional signs of possible CV accident) than lower-cervical manipulations {L-4}. [187]

Risk-management, recommendation 37.   Although the role (alleviating, neutral, exacerbating, causative) of manipulation in CV accidents is unclear, we recommend using a minimal rotation in administering an upper-cervical spine manipulation until better information is available, to maximize the benefit to harm balance.

Risk-management, recommendation 38.   Extrapolating from our recommendation to use a minimal rotation in administering an upper-cervical spine manipulation, we also recommend the use of a minimal rotation in administering any modality of upper-cervical spine treatment.

2.1. Hyperhomocysteinemia as a risk for stroke or TIA

The evidence that elevated plasma homocysteine concentrations (and the associated, reduced plasma concentrations of folate, vitamin B12 or vitamin B6 {L-1b} [192] {L-2c}) [188-191] correlates with an increased risk of stroke is not conclusive {L-5}^GDC – although some evidence directly suggests that specifically sub-optimal levels of vitamin B6, and not hyperhomocysteinemia, are associated with hypothesized increases in the risk of stroke or TIA {L-5}. [193] Although the evidence is inconclusive, this suggests that patients presenting with known hyperhomocysteinemia should be treated with caution {L-5},^GDC but not that an assay of homocysteine levels should be part of an assessment with healthy patients.


2.2. Stroke; predictors of this adverse event subsequent to manipulation

The link between rotation manipulation and stroke is tenuous (above), and at least one study [194] has reported that the maximal angular difference between the head and neck or trunk in the course of a rotation manipulation does not exceed the active physiologic range of motion {L-4}.

However, we concluded that the risk factors for stroke should be accounted for in best practice patterns (Section 5). See Tables 3a to 3h for all identified risk factors that are absolute contraindications, Table 5 for all identified factors that require treatment caution, and Table 6 for stroke-specific risk factors that require treatment modality modification {L-5}. Section 5.3.2 clarifies the clinical meaning of modality modification.