The need for taking routine X-rays prior to manipulative therapy, in particular before vertebral manipulation, was discussed during a session, organized by SOFMMOO, at the 16th Spine Update, in June 2003. A survey before the Panel discussion had shown much uncertainty among practitioners of manual medicine.

In the standard textbooks of manual medicine, this question is not directly addressed. On the other hand, all the international guidelines are strongly against the routine use of X-rays of the spine, and state that radiography under these conditions is indicated only if there are warning signs (“red flags”). The clinical manifestations that come under this heading are shown in the box. They must be searched for with meticulous care, as part of the medical work-up of the patient. The guidelines derived from their presence are supported by research-based evidence, and are justified by the fact that X-rays are expensive, that the radiation burden to the patient is not negligible, and that the demonstration of insignificant or barely significant abnormalities may adversely affect the patient’s view of his or her pain, and may turn him or her into a chronic pain patient. It must be borne in mind that the guidelines deal with diagnostic radiography, not with X-rays taken to assess the risk of manipulation (although, as we shall see, these two concepts overlap), and that they concern the lumbar spine, without any direct reference to the cervical spine.

On the other hand, in its guidelines for the use of imaging techniques in low-back pain cases, ANAES states that “outside the context of looking for evidence of symptomatic low-back pain, the use of imaging techniques is not indicated in the first seven weeks of the low-back pain, unless the envisaged therapeutic modalities (such as manipulation or infiltration) make it necessary formally to exclude any specific form of low-back pain.” Also, past verdicts and settlements have shown that in cases of post-manipulation complications, the absence of X-rays prior to manipulation is regarded, by the experts, as failure to conform to the standard of care (malpractice), even if prior X-rays could not, under any circumstances, have prevented the occurrence of the complications.

There are, thus, two possible attitudes. The first would be to say that proper history-taking and clinical examination should suffice to alert the practitioner to the presence of a red flag, and suggest when an X-ray should be taken (or any other appropriate imaging technique performed), and when there is no need for such an investigation. This would allow radiography costs to be saved, reduce the radiation burden to the patient, and avoid the adverse consequences that may occur when very minor abnormalities are detected. The other attitude is the exact opposite: it would be to say that routinely performed X-rays would allow abnormal conditions to be detected in patients who have no specific signs or symptoms, and whose back pain is considered as “ordinary”, whereas, in fact, the condition is severe and advanced enough to put the vertebral bodies and the spinal cord at risk of serious complications in the event of manipulation. This attitude would also allow the practitioner to comply with the (unwritten but very obvious) rule that everything should be done to avoid a malpractice suit.

There was, therefore, a need for considering this question in greater depth. The procedure adopted by SOFMMOO was as follows: The five papers given at the 16th Spine Update, as well as the comments by members of the Society, were used as a starting point for discussions. The text produced on this site was written by J. Y. Maigne, who had initiated the exercise; and reviewed and revised by Drs G. Berlinson, F. Dumont, J. C. Goussard, M. Marty and P. Vautravers. The lumbar, thoracic, and cervical regions of the spine, and the possible post-manipulation complications at these levels, were considered separately, in the search for specific guidelines regarding the utility of plain X-rays. The definition of manipulation adopted for this exercise was the one given by R. Maigne: “a single, short, sharp thrust that takes a motion segment beyond its normal end-range of motion, but not beyond its anatomical range. It is usually associated with a cracking noise.” Thus, by definition, all mobilization techniques, stretching, extension, massage, and muscle energy techniques (contract-relax) were excluded.

Cauda equina syndrome

• Cauda equina syndrome following manipulation of a herniated intervertebral disc
Cauda equina syndrome is the most serious complication of lumbar-spine manipulation. A 1992 US paper (Haldeman & Rubinstein) found and studied 29 cases in the literature since 1911. Of these, 28 were for sciatica caused by a herniated disc. Sixteen of the patients with this condition had been manipulated under general anaesthesia. The mechanism involved may be a sudden increase in the volume of the herniated disc, or an interruption of the blood supply to the conus terminalis (Balblanc et al).
Two points should be borne in mind. Firstly, manipulation under general anaesthesia has never been practised in France (and is, to the best of our knowledge, an obsolete procedure); it carries a high risk (since the patient is unconscious and cannot react). Secondly, radiography is of no use in the diagnosis of a herniated disc, and could not, therefore, by itself have prevented these complications.

• Cauda equina syndrome following manipulation of a lumbar stenosis
In the paper cited above, the 29th case was that of an achondroplastic dwarf who complained of sciatic pain. In achondroplasia, the spinal canal is stenosed over an extensive distance; this is a contraindication to manipulative therapy. In cases with localized stenosis, no adverse events have been reported in the literature. However, this subject has not been properly studied, and patients with manifestations of spinal stenosis are not, therefore, suitable candidates for manipulation.

• Cauda equina syndrome resulting from the collapse of a tumour-bearing vertebra
Within the context of a wider study, Dupeyron et al recently reported four cases of cauda equina syndrome, of which two appeared to be associated with the collapse of a tumour-bearing vertebra or with epidural spread of a cancer. Since this was a retrospective study, there were no clinical details available; in particular, there was no information on the symptoms that had made the patients seek manipulative therapy. Two papers describe the inappropriate use of manipulation in patients with cancers of the spine. The first concerns a female patient who was undergoing chemotherapy for lymphoma, and who was treated with manipulation of the sacro-iliac joints. This resulted in the collapse of a lumbar vertebra, without any neurological sequelae (Maigne & Lefort). In the second paper, there were no complications, and – curiously – the patients obtained temporary pain relief. To the authors, this suggested an analgesic effect of manipulations, which could at least partly account for the effect of manipulative therapy in nonspecific low-back pain (Vautravers & Lecoq). We have not, however, found any paper (other than that be Dupeyron et al) reporting a cauda equina syndrome following the manipulation of a tumour-bearing vertebra.
The question is, therefore, whether it is conceivable that a vertebra affected by a malignancy to the point where it will collapse with manipulation could produce no red flags? The absence, amidst the plethora of reports of post-manipulation complications, of any published cases of this kind would suggest that this pattern does not exist.

• Cauda equina syndrome as a result of the collapse of an osteoporotic vertebra
Osteoporosis is an asymptomatic condition which, at an advanced stage, may be complicated by vertebral collapse. The received wisdom is that this collapse is never associated with neurological complications; however, this is contradicted by some cases reported recently (Benoist). To date, none of these cases have involved patients who had undergone manipulation. Also, patients over 50 and those with “potential” osteoporosis (e.g. history of steroid therapy) should be routinely X-rayed, according to the guidelines.
Radiography is insufficiently reliable for the diagnosis of osteoporosis, and is not, therefore, considered to be of use. The gold standard is osteodensitometry.

• Cauda equina syndrome as a result of the collapse of a benign-tumour bearing vertebra
Benign tumours of the spine are rare, and usually occur in specific contexts. Osteochondroma and eosinophilic granuloma are seen in children under the age of 10. Giant-cell tumours. osteoid osteoma, and osteoblastoma do not cause fractures. However, complicated active angioma and aneurysmal bone cysts may lead to vertebral collapse with nerve root or cord compromise. The former tends to occur mainly in the thoracic spine, and will be discussed in the relevant section below. The latter produces mechanical back pain, and is seen chiefly in children and in young adults (age bracket 5–25 years). It affects all parts of the spine with equal frequency, causing mechanical pain and, in about 10% of the patients, fractures which may be associated with paraplegia (Papagelopooulos et al). As stated in the guidelines (“patients under 20”), pain in a young patient should be seen as an indication for X-rays. The guidelines do not mention the 20–25-year age group. The occurrence, in a patient of that age, of an aneurysmal bone cyst mimicking recent-onset nonspecific low-back pain is probably extremely rare. However, it would be wise to routinely request X-rays also in these patients, in order to be on the safe side. We would, therefore, plead in favour of an extension of the age bracket envisaged in the guidelines, to ensure that patients aged between 20 and 25 years are also routinely X-rayed prior to manipulation.

Simple vertebral fracture

An uncomplicated vertebral collapse secondary to manipulation is theoretically possible, especially in osteoporotic patients; however, there are no reports to this effect in the literature. It is also possible that a patient who has recently suffered a painful vertebral collapse (either spontaneously or as a result of an RTA) would receive manipulation for what is diagnosed as benign acute low-back pain of recent sudden onset (post-traumatic or not), without any prior X-rays being taken. This scenario would be all the more likely in a female patient in the osteoporosis-prone age group. This rule of good clinical practice should be sufficient.

Spondylolisthesis as a result of pars fracture

The question is whether spondylolisthesis (SPL) as a result of a pars fracture can be aggravated (clinically or anatomically) by inappropriate manipulation. There is no easy answer, because of the lack of data in the literature. However, the question is important because SPL can be detected only by lumbar-spine radiography (except for the rare cases where the displacement of the spinous processes provides conclusive information). There are three patterns: pars fracture, low-grade SPL, and Grade II or higher SPL.

• At the spondylolysis stage
A pars fracture that produces clinical symptoms cannot be missed. There will be pain of comparatively sudden or rapid onset, in an adolescent, frequently following an athletic movement that involves pronounced extension of the lumbar spine. This pattern is pathognomonic, and provides sufficient reason for requesting X-rays. Practitioners would do well to remember what happened in the much-publicized case of an under-age female athlete who complained, during a training session, of sudden low-back pain. Her trainer more or less forced her to go on. The pain subsequently became worse, and then chronic, and the girl’s family sued. The context was not, of course, one of vertebral manipulation; however, the parallels are obvious. Pain, for however short a time, in a young subject is a red flag recognized by the guidelines, which makes X-rays mandatory.

• Low-grade SPL
To the best of our knowledge, there are no cases in the literature of SPL being aggravated clinically or anatomically following manipulation; neither are there any written rules as to the correct procedure to adopt. The personal experience of a large number of practitioners of manipulative therapy would suggest that manipulation not only does not do any harm, but that it is actually beneficial: patients with low-grade SPL have been successfully treated with lumbar-spine manipulation (Young & Koning). These patients had mainly suffered from chronic low-back pain, and had, therefore, been X-rayed (in compliance with the international guidelines according to which X-rays should be taken where the pain has been going on for more than two months).

• Grade II or higher SPL
SPL of grade two and above can give rise to chronic low-back pain or sciatica. Manipulation may not, on the face of it, appear to be the most logical treatment method; however, it is well to remember that, as far as we know, there are no published cases of spondylolisthesis being made lastingly worse by manipulation. The pain mechanism is complex. Lumbar-spine manipulation may work on muscle contraction, and the low-back pain may also be caused by what is happening at the thoracolumbar junction, a region that lends itself to manipulation.

Aggravation of low-back pain or sciatica after manipulation, in the absence of significant X-ray findings

Following manipulation, low-back pain or sciatica may become worse, in the absence of any significant X-ray findings (other than degenerative disc disease, which is notoriously uncorrelated with the clinical manifestations; or mild and long-standing Scheuermann’s disease). This aggravation is usually of short (a few days’) duration, and is followed by a complete return to the former condition; however, in very rare cases it may persist, and patients have been known to sue. Regardless of the underlying mechanism, the fact that there are no (or only very commonplace) visible bony lesions means that X-rays would not have been contributive. A discussion of how such situations could be prevented is beyond the scope of this article; however, it cannot be overemphasized that practitioners should listen to their patients, and try to understand what it is that is making them seek treatment. Some patients do not want to be manipulated, and the practitioner has to accept that. Above all, he or she must identify as accurately as possible where the patient’s pain is mainly coming from, as well as any factors contributing to the chronic nature of the pain (disc, facet joints, pain pathway dysfunction, litigation, etc.). Manipulation should be performed only in cases of strictly mechanical pain, without any inflammatory component. This diagnostic work-up requires meticulous history-taking and a careful clinical examination. X-rays can be useful, but should not be resorted to as a routine investigation. Here, too, the guidelines appear to be sufficient (routine X-rays in patients with chronic pain); however, it is important to realize that X-rays cannot prevent the type of situation discussed here.

Furthermore, there are situations in which radiography would not be justified on medical grounds, but where the patient is very keen to have X-rays taken. Requesting such “off-guideline” X-rays may serve to reassure the patient, enhance the doctor-patient relationship, and prevent the type of situation discussed here. While such patient-driven situations are not covered by the guidelines, they do occur and should be taken into account.

Lumbar spine – conclusions

The guidelines appear to cover all the situations where there is a risk to the spine following lumbar manipulation, and should, therefore, constitute an adequate decision-making aid for the practitioner. There are only two qualifications to this statement. Firstly, in subjects between 20 and 25 years of age, an aneurysmal bone cyst may be present. Since, under these circumstances, manipulation carries a certain risk, subjects in this age group should be routinely X-rayed. Secondly, a patient without any red flags may be anxious to have X-rays done. If manipulation is being considered, it would appear wise to comply with the patient’s wishes, so as to provide reassurance on the condition of his or her spine.
Routine X-rays prior to lumbar manipulation is not evidence-based, and should not, therefore, be performed systematically.

Nerve root and cord compression

The complications seen after thoracic-spine manipulation are the same as those listed above for the lumbar spine. There is no cauda equina syndrome, but cord compression may occur. The only condition not encountered in the lumbar spine, but possibly present at the thoracic level, is complicated active angioma, whose site of predilection (86% of cases) is in the T-spine. The condition is more common in females, especially during pregnancy, and presents as back pain, or as nerve root or cord compression of rapid onset; it may be secondary to trauma, or spontaneous (fracture or epidural haematoma) (Castel et al). The imaging technique of choice is MR.
Thoracic pain is a red flag in some (especially the British) guidelines; patients with thoracic pain should be X-rayed. However, there are guidelines that do not recognize this red flag. The SOFMMOO guideline is that X-rays should be performed only where there are red flags, or where the pain is chronic. In case of doubt, in pregnant women, preference should be given to medical treatment, before considering MRI.

Osteoporotic vertebral collapse

What has been stated above concerning the lumbar spine also applies to the thoracic spine, and will not be repeated here.

Rib fracture

In very rare cases, rib fracture may occur in the course of certain manipulative techniques which put pressure on the sternum and the front of the rib-cage. Elderly patients are particularly at risk, and caution should be exercised when performing such manoeuvres. X-rays cannot be a substitute for this caution.

Vertebral artery dissection

Vertebral artery dissection is the most serious complication that can occur with manipulative therapy. At the present state of our knowledge, this complication is unforeseeable, and the non-utility of radiography is undisputed. Doppler studies appear to be an interesting research approach (Haynes); however, the routine use of Doppler has not been validated to date. The only way to prevent this complication is to follow the SOFMMOO guideline, which states that manipulation involving rotation should not be performed in female patients under the age of 50.

Aggravation of a cervical-spine fracture or sprain

Severe cervical-spine fractures or sprains may occur as a result of trauma. Usually, the forces causing the lesion will be fairly violent; however, a lesser force may result in fractures or sprains, especially in elderly subjects (dens fracture, which may become secondarily displaced) or in osteoporotic patients. If there has been cervical-spine trauma followed by non-remitting pain, X-rays must be taken. The films must show the dens in elderly subjects, and flexion-extension films have to be taken (one week after the accident) to allow any severe sprain with instability to be detected. This requirement is stressed in two recently published papers (Brynin & Yomtob, Crowther).

Aggravation of nerve root pain or neck pain

As with low-back pain and sciatica, neck pain or neck-arm pain may be made temporarily worse by ill-advised manipulation. In the vast majority of cases, everything will return to normal within a short time. However, in rare cases, and under certain clinical conditions, this aggravation may be lasting, and the practitioner may be held liable. In patients with osteoarthritis of the cervical spine or with a high-grade intervertebral foramen stenosis, manipulation may be contraindicated. Since the cervical spine is inherently more susceptible to damage than is the lumbar spine, it is recommended that X-rays be routinely performed prior to manipulative therapy.

One special case: malformations of the occipito-cervical junction

Malformations of the occipito-cervical junction may be bony (basilar impression, block vertebra) and/or nervous (Arnold-Chiari malformation). Plain X-rays cannot, under any circumstances, rule out nervous malformations, for whose diagnosis MRI is essential. Arnold-Chiari malformation is frequently associated with syringomyelia, which clinically often presents as ill-defined neck-arm pain (with diminished triceps reflex). Although X-ray findings and clinical findings do not necessarily agree, and while there are no data in the literature, the possibility of bony malformations is a reason for systematically requesting X-rays prior to manipulation.

An analysis of the complications that may follow manipulation, and the discussion of the utility of pre-manipulation X-rays for the prevention of the various complications, show that the existing guidelines are, overall, sufficient to guard against complications at the lumbar and the thoracic level. The only situation not covered is that of young subjects, between 20 and 25 years of age, in whom an aneurysmal bone cyst is theoretically possible.

In the cervical spine, particular attention must be paid to malformations of the occipito-cervical junction, which may not be detected by clinical examination; and to the presence of advanced degenerative disease. Even though there are no reports of adverse events in the literature, manipulation in such patients may involve a certain risk of complications. This is why SOFMMOO recommends routine X-rays of the cervical spine. AP and lateral dens-centred views are required in patients with headaches of presumed cervical origin or with high cervical spine pain, if manipulative therapy is being considered. This guideline applies even if the patient’s condition is not chronic.

These guidelines are not definitive. They reflect our present state of epidemiological knowledge, and may be modified subsequently.

The guidelines are as follows.

• ANAES. Diagnostic, prise en charge et suivi des malades atteints de lombalgie chronique. 2000;p31.
• Assendelft WJ, Bouter LM, Knipschild PG. Complications of spinal manipulation: a comprehensive review of the literature. J Fam Pract 1996 ;42:475-80
• Balblanc JC, Pretot C, Ziegler F. Vascular complication involving the conus medullaris or cauda equina after vertebral manipulation for an L4-L5 disk herniation. Rev Rhum [Engl Ed]. 1998;65:279-82.
• Benoist M. Complications neurologiques des fractures vertébrales ostéoporotiques. Revue Médecine Vertébrale. 2003;9:7-11.
• Brynin R, Yomtob C. Missed cervical spine fracture: chiropractic implications. JMPT 1999;22:610-4.
• Castel E, Lazennec JY, Chiras J et al. Acute spinal cord compression due to intraspinal bleeding from a vertebral hemangioma: two case-reports. Eur Spine J. 1999;8:244-8.
• Crowther ER. Missed cervical spine fractures: the importance of reviewing radiographs in chiropractic practice. JMPT 1995;18:29-33.
• Dupeyron A, Vautravers P, Lecocq J, Isner-Horobeti ME. [Complications following vertebral manipulation-a survey of a French region physicians]
  Ann Readapt Med Phys 2003;46:33-40
• Goussard JC. Radios systématiques ou non avant manipulation : que nous disent les statistiques d’assurances ? Revue Médecine Vertébrale. 2003;11.
• Haldeman S, Rubinstein SM. Cauda equina syndrome in patients undergoing manipulation of the lumbar spine. Spine. 1992;17:1469-73.
• Haynes MJ. Les artères vertébrales et les mouvements cervicaux : intérêt du doppler artériel en tant qu’examen prémanipulatif. JMPT 2002;25:256-67.
• Lee KP, Carlini WG, McCormick GF, Albers GW. Neurologic complications following chiropractic manipulation: a survey of California neurologists. Neurology 1995;45:1213-5.
• Maigne JY, Lefort S. Tassement lombaire après manipulation non indiquée. Revue Médecine Orthopédique. 1986;4:24.
• Marty M. Les recommandations des "guidelines" sont elles suffisantes ? Revue Médecine Vertébrale. 2003;11.
• Kraft CN, Conrad R, Vahlensieck M et al. Non-cerebrovascular complication in chirotherapy manipulation of the cervical vertebrae. Z Orthop Ihre Grenzgeb 2001;139:8-11
• Papagelopoulos PJ, Currier BL, Shaughnessy WJ et al. Aneurysmal bone cyst of the spine. Management and outcome. Spine. 1998;23:621-8.
• Rousseaux P. Pièges neurologiques des manipulations vus par le neuro-chirurgien. Revue de Médecine Vertébrale. 2002;7:25-29.
• Vautravers P, Lecoq J. Pièges redoutables en rapport avec les manipulations vertébrales. In : Hérisson C, Vautravers P. Les manipulations vertébrales. Paris: Masson, 1994:296-304.
• Young KJ, Koning W. Spondylolysis of L2 in identical twins. J Manipulative Physiol Ther. 2003 ;26:196-201.