By Arthur Croft, DC, MS, MPH, FACOThe Back Letter, a newsletter-like Spine spinoff put out by Lippincott-Raven, recently ran a piece on this continuing imbroglio of whiplash. You might recall that last year they visited this issue again when the Lithuanian paper was published in Lancet.1 Based on the nod that the paper got from one of the authors of the Quebec Task Force on Whiplash-Associated Disorders (WAD), the Back Letter article offered up this Lancet paper as just a little more fuel for the controversy about whether one can be significantly injured in low speed rear-impact crashes (LOSRIC).
Recently, the ACA Journal sent this Back Letter issue to me asking for my take on it. In fact, in the past few months, I'd estimate that at least a dozen other DCs have also faxed copies to me. The original piece made reference to an unpublished research work by an author named Castro. I didn't save it, unfortunately, so I can't cite the exact conclusions that were drawn, but as I recall from reading the article in the Back Letter (which also made reference to a few other papers), the general drift was that here is further evidence that LOSRIC don't typically result in injury -- at least, not significant or lasting injury.
Of course, it's hard to do anything with unpublished material, so I just put the thing aside.
However, the Castro et al. paper has since been published2 and has even been awarded the European Spine Society AcroMed Prize for Spinal Research (whatever that is). Now, with the paper firmly in my hot little hands, I can give you my take on it.
The authors conducted 17 crashes with volunteers -- mostly male -- of an average age of 33 years. Subjects were given a physical examination, computerized motion analysis, and MRIs with contrast. The range of velocity changes for these crashes was 5.4 mph to 8.8 mph, with an average of 7 mph. The range of acceleration of the target (struck) vehicle was reported to be 2.1-3.6 g, with an average of 2.7 g. The only reported abnormality noted was a limit of cervical spine motion in one volunteer lasting for 10 weeks. The study concluded that the "limit of harmlessness for stresses arising from rear-end impacts with regard to the velocity changes lies between 10 and 15 km/h (6.2-9.3 mph)."
In this type of testing, there are several limitations to conclusions that can be drawn. For one thing, it is nearly impossible to dodge the volunteer awareness issue. They do know exactly what is about to happen to them. And, while most crash test researchers believe they dismiss this problem peremptorily by obstructing the subject's view or playing music through headphones to prevent them from hearing, this argument is weak.
For one thing, this superficial camouflage almost certainly does not equate to the degree of nonawareness of real world occupants. Having participated in such tests, I must conclude that the staging area is something like that of a movie set. Between takes, we all stand around, relaxed, talking or taking notes. When the action starts, we all gather at our observation sites, look up, man the cameras, and generally become an alert mob. Anyone« peripheral vision.
Another limitation in this type of research is that the volunteers tend to be screened for preexisting problems: a luxury real world motorists are not afforded. They also tend to be a rather more adventurous lot, with many being the researchers themselves. And, in most cases, the subjects are usually not given adequate long-term followup to determine how they fare in the long run.
Then there is the issue of occupant positioning, which tends to be rather optimal in the majority of these experimental crashes. Moreover, it is not possible to subject the data to statistical analysis, because the actual number of crashes is so small. In the end, it is not possible (nor permissible on scientific grounds) to generalize about the universe of real world motorists from the results of these small tests: the studies lack external validity. Still, the results are useful and very interesting.
As to the computerized motion analysis used in the present study, it was done using some type of ultrasound device. The authors admitted that the significance of this test was not known. The problem is that this is not pointed out until later in the report, and it is not mentioned in the abstract. Using a nonvalidated device to measure outcome is somewhat dubious. While the MRI study is certainly a valid one, I suspect that limitations in resolving power may prevent us from appreciating the smallest lesions resulting from this form of trauma. Nevertheless, the authors did not produce any large lesions.
I might also point out a couple of things that are likely to be used by accident reconstructionists when they manage to get hold of this paper. The velocity changes reported were for the car, not the occupants. In low speed crashes, the coefficient of restitution of most car seats will be from 0.3 to 0.7. That means that the occupant will reach delta Vs higher than the vehicle by a factor of 30-70 percent.
Also, the mean acceleration reported was for the car, not the occupant. We know that the peak acceleration will always be higher than the mean acceleration, and the acceleration of the occupant's head will be several times higher than that of the car. Whether this was a mere oversight, or whether the authors did not consider the occupant's acceleration to be important, is difficult to say. However, in similar speed volunteer crash tests, peak linear head accelerations of 16 g have been reported.3 If I was an AR working for an insurance company, I'd go for 2.7 g.
Perhaps the most egregious fault one can find with this paper is the reporting that 6.2-9.3 mph LOSRIC are "harmless." The authors caused 5 of 17 (29 percent) volunteers to sustain injury, including one with a reported loss of range of motion lasting 10 weeks. Other recent volunteer crash tests have also produced minor injuries in more minor crash tests: 29 percent at 2.5 mph and 38 percent at 5 mph delta Vs.4
Regardless of one's perception of the seriousness of such injuries, one is not entitled to call the crashes that produce them "harmless." Moreover, epidemiological data have shown that, 6.2-12.4 mph.5,6 Meanwhile, also running counter to the claims that injuries cannot occur in whiplash, there is an entire literature of clinical and experimental work which is necessarily and carefully ignored by the naysayers of whiplash. Thus, the author's conclusions are clearly not consistent with the real world.
This pestiferous trickle of literature which attempts to prove the innocence of whiplash-type crashes (LOSRIC) will not stop, I suppose, until we actually devise a way to either prevent crashes altogether, or until we actually make our cars crashworthy at all speeds. In the meantime, we will have to take a careful look at each paper as it comes along in order to determine how the authors have managed to apply their popular fait accompli. The fait accompli, in this case, is the deceptive form of reverse engineering of the well-accepted scientific process which normally follows the order of observation to hypothesis generation to testing of hypothesis and so on, after which we either accept the hypothesis or reject it and form another.
The authors of the "paltritude of whiplash" genre of work first confect the observation, i.e., whiplash does not cause injury (or long-term injury, or significant injury). Then they devise an experiment to be consistent with their "observation." This is not difficult to do, although the watchdogs will always catch them at their mischief.
The Lithuanian whiplash study was flawed to such a degree that the authors were not actually looking at what they thought they were looking at, and the watchdogs noticed.7-9 The Quebec Task Force on WAD drew a few dubious conclusions, and again, the watchdogs were there.10
C'mon guys, get over it: whiplash is real. Try it sometime and you'll be a believer.
Arthur C. Croft, DC, MS, FACO
Director, Spine Research Institute of San Diego
San Diego, California
Click here for more information about Arthur Croft, DC, MS, MPH, FACO.