Contents of this page

"Whiplash"cases from my practice

Whiplash: A Medical Doctor's Review of the Literature

Headaches/Neckaches from "Whiplash Injuries"

What follows is a broad survey of "whiplash." Although I cite here only a small number of the numerous publications on this subject, the ones selected are, in my opinion, particularly telling. Those interested in learning more, can begin to access the literature from my references.

How is the Term Used?

Crowe coined the term whiplash in 1928 in a talk on neck injuries in rear-end automobile collisions in the U.S., but regretted it later (Breck & Van Norman, 1971). In 1969, he said that he "never published the term, but it spread across the country, got into published literature, and became a household word which is used thoughtlessly by laymen, attorneys and physicians. It describes only the manner in which a head was moved suddenly to produce a sprain in the neck."

As a description of a movement, the term is an inaccurate metaphor, for the head and neck do not imitate the movement of a snapped whip. In addition, the original use of the term as a description of head-neck movements from rear impacts is no longer adhered to by many workers. For example, one prominent group of authors (in Australia), in a review of "whiplash injury" in 1994, state that "Although classically described in association with movements in the sagittal plane following rear-end impact,...neck pain may also follow lateral or frontal collisions" (Barnsley, Lord, Bogduk, 1994). So "whiplash movement" is no longer used by most authors in its original metaphoric sense. In fact, most don't use the term to refer to a movement, but use it to refer to a syndrome or an injury. What is this syndrome and what is this injury?


The Early "Whiplash Syndrome"

The symptoms occurring soon after "whiplash injury," from movement of the head in any plane, have been particularly well-studied by Sturzenegger et al. (1994) (in Switzerland). The bar graph, made from their data, shows what percentage of patients had each of the four most-common symptoms. 87% of patients had multiple symptoms. Many of the patients reported disappearance of symptoms within a few days.



Similar immediate post-collision symptoms in 21 patients were recorded by Ettlin et al. (1992) (in Switzerland). These were neck pain in 76%, headache in 67%, nausea in 48%, upper limb pain in 43%, and vertigo in 38%. Note that all studies report neck pain and headache as the most-common symptoms, but vary considerably in their listing of further symptoms, due to differences in interview methods, patient populations, etc.

The symptoms fade away in most patients during the subsequent weeks and months after the collisions, so that only a small percentage of patients remain symptomatic beyond 6 months or so. These patients have the persistent late syndrome. Borchgrevink et al's. (1995) data on 52 patients in Norway shows that symptom prevalence fell from 2 weeks to 12 months and then remained static for the next year.


The Late "Whiplash Syndrome"

Radanov et al. (1995) did a prospective study of the symptoms persisting 2 years after "whiplash injury" in 117 patients in Switzerland. They did not limit cases to those with rear-end crashes. The bar graph below, made from their data, shows the prevalence of symptoms in their series 2 years after the patients' collisions.

Because the patients who remained symptomatic at 2 years had "a higher intensity of initial neck pain and headache and a significantly increased prevalence of radicular deficit...," the authors felt that the long-term prognosis is probably related to the severity of trauma. However, they did not document these radicular deficits nor provide evidence linking the intensity of initial neck pain to traumatic severity. Moreover, they did not consider that more intense initial neck pain and headache might be due instead to patients' reactions to their traumatic events.

In contrast to the conclusions of Radanov et al., Karlsborg et al. (1997), from a prospective study of 39 patients in Denmark, linked the development of the late syndrome to the presence of stressful events in patients' lives at the time of or soon after the automotive collisions. (For the detailed psychological evaluation of 29 of their patients, see Smed, 1997.


The Headaches of the Late Syndrome Resemble the Chronic Post-traumatic Syndrome

Only Foletti and Regli (1995) (in Switzerland) have classified the headaches in patients with the ate whiplash syndrome. They appropriately used International Headache Society criteria. 74% of their 47 patients had tension-type headache, 15% had migraine and 11% had cervicogenic headache. These findings are very close to what I found in my patients with chronic post-traumatic headaches (Haas, 1996).

Samples of patients with the "whiplash syndrome" have a higher incidence of neckache than those with the chronic post-traumatic (post-concussional) syndrome (Cartlidge and Shaw, 1981), but otherwise the clinical picture in these two groups is identical. Most patients I see for headache after automotive collisions have the same set of symptoms as those I see for headache after head trauma.


The Late Syndrome in Lithuania, Singapore, Greece, and Germany

Two overlapping groups of investigators studied the symptoms that Lithuanians developed after rear-end car crashes, because Lithuanians are unaware, in general, of the chronic whiplash syndrome and therefore have no expectation of developing chronic symptoms after such crashes. The first study, by Schrader, Obelieniene, Bovim, et al. (1996) was a retrospective, questionnaire-based comparison of symptoms in victims of rear-end collisions with those in matched controls. The collisions required police attention. 172 drivers, mostly men, plus 30 women passengers, completed the questionaires and so did 202 controls. None of the accident victims had persistent or disabling symptoms from the accident, but 31 recalled having had acute transient neck pain.

The next study, by Obelieniene, Bovim, Schrader, et al. (1999) was a prospective evaluation of rear-end crash victims in the same city. 210 victims were consecutively identified from the daily police records. Neck pain and headache were evaluated by mailed questionnaires shortly after the accident and 2 months and 1 year later. 210 sex and age matched controls were randomly taken from the population register of the same geographical area and evaluated for the same symptoms immediately after their identification and at 1 year. Initial pain was reported by 47% of accident victims; 10% had neck pain alone, 18% had neck pain together with headache, and 19% had headache alone. The median duration of the initial neck pain was 3 days and the maximal duration was 17 days. The median duration of headache was 4.5 hours and the maximum duration was 20 days. At 1 year, there were no significant symptomatic differences between the accident victims and the controls. The authors concluded that without an expectation of developing chronic pain from rear-end collisions, without involvement of the therapeutic community and insurance companies, and without litigation, symptoms after an acute "whiplash" injury are self limiting, brief, and do not evolve into the so-called "late whiplash syndrome."

Balla (1982) looked for the late whiplash syndrome in Singapore by interviewing all neurologists and neurosurgeons as well as some orthopedists and psychiatrists there. "But for the odd case seen in those with European backgrounds the Late Whiplash Syndrome was unheard of." He also reviewed 20 patients seen in an Orthopedic Clinic more than 2 years before for "acute whiplash injury." Their initial head and neck pains were the same as those of Australians, but none had developed symptoms of the late syndrome.

A prospective study of the outcome of acute whiplash injuries in Greece by Partheni M, Constantoyannis C, Ferrari R, et al. (2000) consecutively recruited 180 accident victims from the emergency ward. A standard questionnaire asked about the usual symptoms. Subjects were followed for 6 months. In the initial 4 weeks after the accident, subjects reported neck pain, headache, shoulder pain, arm numbness or pain, and dizziness, but at 4 weeks more than 90% had recovered from these symptoms, while the remainder had minor symptoms not requiring therapy. No patients had chronic disability. The authors concluded that in Greece, symptoms after an acute whiplash injury are self-limiting, brief, and do not appear to evolve into the so-called late whiplash syndrome.

The chronic whiplash syndrome is also remarkably uncommon in Germany, according to two studies mentioned in a review of late whiplash syndrome by Ferrari and Schrader (2001).


Effects of Litigation

In the U.S.A. and other industrialized countries many but by no means all victims of car crashes retain lawyers to sue for monetary compensation for their symptoms. For example, among 14 consecutive patients with chronic head or neck symptoms after car crashes that I've recently seen, 10 were in litigation.

A recent epidemiological study of the effects of litigation on closure of claims for injuries after car crashes was carried out in Saskatchewan, Canada by Cassidy et al. (2000). Closure and resolution of symptoms (which are highly correlated) was determined for a period when a tort system was in place and for a subsequent period when it was replaced by a no-fault system, wherein victims could no longer sue for pain and suffering. The findings were that after the changeover the incidence of "whiplash" claims dropped by 28% and the median time to the closure of claims was reduced by more than 200 days.


Evidence, or Lack of Evidence, For Injury To the Neck

Evidence From Radiographs and MRI

Borchgrevink et al. (1995), in Norway, obtained radiographic and MR images of the necks of 52 "whiplash" patients within 4 days of their collisions.

Radiographic Findings, Listed by Number of Patients:

MRI Findings, Listed by Number of Patients:

These findings, they said, are common in asymptomatic persons. Curiously, those with lordosis loss on the radiographs (but not the MRIs) had, as a group, fewer symptoms than the other groups. The group with spondylosis had significantly more symptoms than the other groups. Although the authors attribute this to the spondylosis itself, it could as well be from other factors within this older group of patients. The authors concluded that MRI is not beneficial in the common hyperextension-flexion injuries. I know of no study which contradicts this conclusion.

Karlsborg et al. (1997), in Denmark, did cervical radiographs and MRIs on 39 patients with "whiplash injuries" without neck fractures or direct head trauma. All the radiographs, done soon after admission to the Casualty Department, were without fractures (an exclusion criterion), dislocations, or postural abnormalities. Among the 30 patients who had MRIs (at a median interval after the accident of 33 days), none showed signs of edema or hemorrhage in the spinal cord or peri-spinal tissues, 7 had minor abnormalities considered nonspecific or due to pre-existing disease, and 3 had disk protrusions or herniations without correlation to the clinical findings.

These studies show that although severe cervical injuries have been well-documented after major automobile collisions (rips in the anterior longitudinal ligament, fractures of the vertebral end plate, separations of the disk from the end plate, and disk herniations), such injuries are not present after the usual collisions which trigger the typical "whiplash syndrome."


Evidence For or Against Pain of Cervical Origin

The concept that chronic post-"whiplash" pain is from neck abnormalities, specifically from traumatic abnormalities in the zygapophyseal joints (facet joints of the lateral articular pillars), has been forcefully advocated by certain Australian workers, whose papers I review in the following sections.


Can the Cervical Facet Joints Generate Pain?

To define the particular generation of pain by these joints, Dwyer, Aprill, and Bogduk (1990) injected those of 5 volunteers with radiographic contrast solution. The referred pain from each joint varied little among the subjects. For the C2-3 joint it was from the lower occiput to the 3rd spinous process. For the C6-7 joint it was from the lower postrolateral neck to the lower scapula. The intermediate joints gave distributions between these two. The pain was always limited to the side injected.


Are the Cervical Facet Joints the Source of the Neck/Head Pain in Chronic "Whiplash" Syndrome?

Anesthetic Blockades of Cervical Nerves and Joints

In an early approach to this question, Bogduk & Marsland (1986) blocked the third occipital nerve (supplies the C2-3 facet joint) with a local-anesthetic in 10 consecutive patients with chronic, continuous "occipital or suboccipital headache," of which 3 were sequelae of motor-vehicle accidents (MVAs). 7 of the 10 patients, including the 3 with MVAs, had complete pain relief for the duration of the anesthetic's action, but no relief from injections of other nerves. Radiographs showed no arthropathy of the C2-3 joints. Nevertheless, the authors favored "traumatic arthropathy or degenerative joint disease" as the "causative pathology."

Barnsley, Lord, & Bogduk "explored the utility of comparative local anaesthetic blocks in the diagnosis of neck pain stemming from the cervical zygapophyseal joints." 47 patients with "chronic neck pain following whiplash injury" attributed to a motor-vehicle accident were studied. The initial joint blocked for each patient was that which best fit the patient's pain location, according to previously determined pain distribution patterns. The initial block was with either short-acting lignocaine (lidocaine) or long-acting bupivacaine, chosen randomly and blinded to injector and patient. When definite or complete loss of pain was produced by a block, the same joint was blocked by the other anesthetic two or more weeks later. When the initial block was negative, the next joint most consistent with the patient's pain distribution was blocked. The assumption was that only patients with genuine joint pain would have longer pain relief from bupivacaine than lignocaine.

44 of the 47 patients had two positive blocks at a single joint, and 34 of these patients had longer relief with bupivacaine. Because this difference was unlikely to have been due to chance, the authors concluded that their comparative blocks identified true zygapohyseal joint pain in these 34 patients. No clinical details of the patients' pain were given and headache was not mentioned.

In 1995, Barnsley, Lord, Wallis, and Bogduk followed the above report with a study of the prevalence of "chronic cervical zygapophyseal joint pain" in patients with "chronic neck pain after whiplash injury." They did the same comparative blocks as before. 38 of the 50 selected patients completed the investigation. Curiously, the listed duration of the neck pain was identical to that in the above study: a mean of "54 months." Moreover, the radiographic reproduction of needles placed for an anesthetic injection was the same in this paper as in the preceding one.

They concluded that of the 38 patients who completed the study, "27 unequivocally met the predetermined criteria for cervical zygapophyseal joint pain." These were that complete pain relief would occur when the joint was blocked with both lignocaine and bupivacaine, and that it would last longer after the latter. 12 of the 27 patients met these criteria from blockage of the third occipital nerve.

Lord, Barnsley, Wallis, and Bogduk (1994) did a more extensive investigation of blocking the 3rd occipital nerve in patients referred for the study and therapy of "neck pain after whiplash injury" (see Bogduk & Marsland, 1986, above). 29 of 100 consecutive referrals were not included in the study, because they lacked headache (they were investigating "third occipital nerve headache"). 16 others withdrew for various reasons, leaving 55 investigated patients. They were injected with the same two anesthetics used in prior studies during two different sessions at the site of one 3rd occipital nerve. Pain relief lasted longer after the bupivacaine than after the lignocaine blocks in 27 of the 55 patients. The authors interpreted this pattern of relief to mean that the 3rd occipital nerve was the source of painful impulses responsible for the patients' headaches.

Curiously, "there were no notable differences between those patients who responded to 3rd occipital nerve blocks and those who did not, with respect to demographic features, psychological profiles, mode of accident, and the duration, frequency, severity, and quality of their pain." However, a much higher percentage of responders had headache that overshadowed their neck pain and had tenderness over the C2-3 joint. 14 of 25 patients whose pain was unaffected by blocking the 3rd occipital nerve got pain relief from blocks of lower zygapophyseal joints.

The authors provided no details on the patients' headaches or neck pain. They did not say whether pain was unilateral or bilateral.

The authors stated that the mechanism of third occipital nerve headache has not been determined, but favored pain originating from injured zygaphophyseal joints (the study did not include CT or MR scans of the joints).


Anesthetic vs. Corticosteroid Injections of Cervical Zygapophyseal Joints

Barnsley, Lord, Wallis, and Bogduk (1994) injected either a depot corticosteroid (betamethasone) or a local anesthetic (bupivacaine) into a single cervical zygapophyseal joint identified as a source of pain by differential nerve blocks in patients referred to their unit for chronic neck pain "attributed to a motor-vehicle accident." 21 patients received the corticosteroid and 20 the anesthetic. The result was that the two groups did not differ significantly in duration of pain relief. "In over half the patients, neither treatment provided relief of pain for more than a week..." Thus, the authors counseled against using intra-articular steroid injections for "cervical zygapophyseal joint pain after a whiplash injury."

As in the above studies, no information on the features of the patients' pain is given and "headache" is not mentioned. The authors thought that a placebo response was an unlikely explanation for the instances of long-lasting relief from either agent. They suggested that the relief obtained from either injection substance may have been from "stretching of the joint capsule."

Radio-frequency Neurotomy Therapy

Two years after the above report of the therapeutic ineffectiveness of facet-joint steroid injections, Lord, Barnsley, Wallis, McDonald, & Bogduk (1996) reported that percutaneous radio-frequency neurotomy of the nerves supplying the "painful joint" (quotes are mine) can give long-lasting relief from chronic neck pain following whiplash injury. Of 54 screened patients, 24 met the criteria for inclusion in the study, by reporting longer pain relief after blocks of a facet joint with long-acting bupivacaine than with short-acting lidocaine and no relief from saline injections. "In 17 patients the neck pain was predominantly unilateral, stemming from one zygapophyseal joint; the other 7...had more than one source of pain." The treated joints were from C3-4 through C6-7. 12 patients received a radio-frequency neurotomy (the nerves to the joint were denatured) and 12 others (the controls) had needle placement without activation of the denaturing current. The operator was blinded to these two procedural differences and patients were blinded by regional anesthesia which blocked sensations of heat or pain during the radio-frequency denaturation of the nerves.

RESULTS: "By 27 weeks, 1 patient in the control group and 7 in the active-treatment group remained free of pain."

In the accompanying editorial, North dissented from the authors' belief that temporary relief of pain after blocking nerves with local anesthetics or denaturing them with radio-frequency unequivocally identifies the source of the pain (attributed to the facet joints in this and the prior studies of this Australian group). His opinion was based in part on his work on the specificity of diagnostic nerve blocks for sciatica due to lumbosacral spine disease. His explanation for the non-specific relief is that local anesthetic blockade "can be effective when it is directed to areas of referred pain." He mentioned that "Even central pain...can be relieved by distal local-anesthetic blocks in the area or segment where the pain is perceived or projected."

Sapir & Gorup (2001) later compared the effect of radiofrequency neurotomy on "whiplash" patients in litigation with those not in litigation. The subjects selected were limited to those with only cervical complaints. The two-phase diagnostic anesthetic blocks of the Australian group were used to select the joint for nerve destruction. Of 60 patients fulfilling entry criteria, 46 completed the study: 28 litigants, 18 non-litigants. Clinically significant and comparable reductions in pain were reported by both groups up to the end of the study one year after treatment. Clinical details were not given and North's views on facet-joint blocks were not addressed.

Trigger-point Injections With Water

Byrn et al. (1993) compared the effects of subcutaneous injections of water with those of saline into "tenderpoints" and "triggerpoints" (T-points) in the neck and shoulders of patients with long-standing "whiplash syndrome" refractory to prior therapies. The water- and saline-injected groups had 20 patients each. Though the ampules containing the fluids were unlabeled, the injectors were not well-blinded, because the water injections stung like a wasp while the saline ones did not. Patients were injected at from 1 to 3 sessions during the first 2 months of this 8-month trial. Patients received from 2 to 116 injections per session, based on the number of T-points found. The water injections were significantly more effective (reduced pain more) than the saline ones after the end of the first session and at the 3 and 8 months follow ups, but not at the 1-month follow up.

SPECT and PET Images of the Brain

SPECT (single photon emission computed tomography)is considered more sensitive than MRI for the detection of brain injury. However, not all SPECT abnormalities can be attributed to brain injury, since they are also seen in patients with painful states and mental abnormalities, such as depression. Bicik et al. (1998) found SPECT and PET (positron emission tomography) "of doubtful value in the routine evaluation of late whiplash syndrome" after studying 13 patients with "persistent head or neck pain and cognitive deficits" but normal CTs or MRIs 6-81 months after "whiplash injury." Their scans were compared to those of controls without neurologic disorders (comparison with those of chronic tension-type headache patients would have been better, in my opinion). The main finding was that, as a group, the patients' PET scans showed decreases of fluorodeoxyglucose in the frontopolar and anterior temporal cortices and in the putamen (meaning lowered metabolism). However, these tests had low sensitivity, for only 8 of 13 patients showed any of the metabolic abnormalities. The authors felt that these changes were not necessarily due to brain damage, because the putamen is not very susceptible to traumatic damage, because others had identified similar abnormalities in depressed patients, and because their study found a correlation between the frontopolar changes and scores for depression.


Effect of Laboratory Controlled Rear-end Automobile Colisions on Human Subjects

Two recent studies have provided quantitative information on volunteers' bodily motions during rear-end car crashes staged in laboratories and on their symptoms afterwards.

The First Study

Castro et al. (1997) used biomechanical, kinematic, and other techniques to study the effects of 17 rear-end automobile collisions and 3 rear-end bumper-car (amusement-park car) impacts (with anticipation neutralized) on 14 male and 5 female volunteers.

The speed of the bullet (impacting) cars averaged 21 km/h and the velocity changes of the target cars ranged from 8.7 to 14.2 km/h. Mean acceleration of the targets was from 2.1 g to 3.6 g. These values are comparable to those of the majority of cars crashed into from the rear on the road, for in 65% of rear-end collisions productive of "injuries," the velocity change of the struck vehicle was no higher than 15 km/h, according to the authors. The impact speed of the bullet bumper car averaged 12.2 km/h and the change in velocity of its target car ranged from 8.3 to 10.6 km/h--values close to those for the target automobiles. Mean acceleration of the target bumper cars ranged from 1.8 g to 2.6 g.

At the moment the automobile was impacted, the lower seat upholstery of the backrest deformed at the passenger's pelvis, which then moved forward, followed by forward movement of chest and shoulders. At this point, the body was straightened, but the head was still in its resting position. Then the head extended backwards in relation to the upper body, and maximal extension was reached when the head contacted the headrest. The extension angle between the head and upper body varied from 10 to 47 (mean, 20) degrees. This is not hyperextension. The motion pattern recorded in the bumper-car crashes was comparable to that in the automobile collisions, but in the former the head hyperextended, because there was no head restraint. The maximal recorded extension in the 3 bumper car crashes was 80 degrees.

At the examinations one day after the crashes, 5 of the 17 subjects in the automobiles reported mild headache or cervical discomfort appearing shortly afterwards. These symptoms disappeared within 7 days in all. None of the 3 subjects in the bumper cars experienced symptoms. Cervical MRIs on the day after the crashes were identical to the pre-crash MRIs.


The Second Study

Brault et al. (1998) subjected 21 women and 21 men seated in front passenger seats of stationary cars to rear-end car crashes (without immediate forwarning) at speeds of 5 and 10 km/h on separate days. These crashes moved the impacted cars to speeds of 4 and 8 km/h (less than in the above study). After each of the two collisions, subjects were re-examined within 30 minutes, on the next day, and then until symptoms and signs resolved.

Nearly all the reported symptoms began after the subjects left the test site. A minority of the subjects reported posterior neck symptoms and headache on the following day. The higher velocity crashes produced no more symptoms than the lower ones. The median duration of all symptoms after the higher velocity crashes was 24 hours. No difference in symptoms was found between the men and the women.


Summary and Interpretation of the Data on "Whiplash"

The term whiplash is a misleading metaphor for the head and neck movements occurring during automobile crashes and as such hinders scientific inquiry into the head and neck symptoms that follow them.

The common early transient symptoms after car crashes probably have a physical basis, for they occur in volunteers subjected to crashes in the laboratory and in accident victims in countries unaware of the dangers of "whiplash." The less common, often-disabling, chronic symptoms of headache, neckache (generally both), and mental and emotional disturbances are probably not caused by physical injury to the head or neck. Supporting this view is the absence of identification of injured structures, the absence of chronic symptoms in certain countries, and their absence in volunteers subjected to rear car crashes.

The strongest evidence for physical injury as the basis of the chronic symptoms is the evidence for symptomatic relief from anesthetic blocks of the cervical facet joints and radio-frequency denaturation of their nervous innervation. However, the advocates of this view report that a considerable proportion of the patients referred to them do not meet their criteria for painful facet joints. So, the facet-joint hypothesis can not be a unified hypothesis for the chronic "whiplash" syndrome.

I don't favor this hypothesis for several reasons: The criticisms raised by North about the interpretation of positive responses to nerve blocks seem important to me. The lack of identification of facet-joint abnormalities on radiographs and CT scans seems contradictory. Shouldn't damaged joints show traumatic changes eventually? In addition, shouldn't they eventually heal if the traumatic damage is not severe enough to produce radiologic changes? Most of my patients' headaches and neckaches are not compatible with cervicogenic headache, in that they are generally bilateral, often with frontal predominance, and unaffected by cervical movements. The dearth of clinical information on patients said to have facet-joint pain prevents a comparison of these patients with mine. Patients with advanced cervical arthritis do not suffer as do those with "facet-joint pain." This doesn't make sense to me. In addition the "whiplash" patients' disabilities and many mental and emotional symptoms seem disproportionate to disease in a cervical joint.

A desire for monetary remuneration through litigation is an unsatisfactory general hypothesis for the chronic syndrome, since many patients do not litigate. However, litigation seems to contribute to symptom prolongation (see Cassidy et al. above), whether by encouraging patients to exaggerate symptoms, by the strain put upon them by the legal process, or by other means.

I'm impressed by the resemblance of the chronic "whiplash" syndrome to the chronic post-traumatic (post-concussive) syndrome, which is as controversial and poorly understood as the "whiplash" syndrome. I favor their being abnormal mental and emotional states triggered by particular traumatic events. I'm working on developing this viewpoint and the data which support it.

The chronic "whiplash syndrome" responds poorly to therapies in the great majority of patients, as does the chronic post-traumatic syndrome. In my experience, a fair percentage gradually improve and some become asymptomatic over the course of a year or more in relation to medicines such as amitriptyline and dextroamphetamine along with a gradual re-organization of their lives.

Home page | Post-traumatic Headache Diagnosis | Post-traumatic Headache Therapy | Top of page
Last updated on Wednesday, April 10, 2002
By David C Haas, MD; e-mail address: see Home page
Hosted by SUNY Upstate Medical University