To the Editor: As an osteopathic family physician
who incorporates osteopathic manipulation therapy (OMT) into my daily practice,
I was concerned by the design and conclusions of the article by Drs Bove and
Nilsson.1 The article should lead one to
conclude only that the particular method of manipulation used was not helpful.
Before dismissing the role of manipulation from the therapeutic armamentarium
for headaches, it is important to look more closely at other relevant factors.
First, OMT is consistently more comprehensive than the treatment described
in this study (techniques were confined to the cervical spine and musculature
surrounding the cervical and upper thoracic spine). Holistic therapy mandates
that one treats the whole body, not just the affected part. By so doing, OMT
is often successful in reducing symptoms, improving outcomes, decreasing direct
and indirect costs (such as physician office visits), as well as decreasing
the frequency of treatments.2
Finally, the outcomes measured involved symptom assessment over 19 weeks.
I would suggest that treatments for episodic tension-type headaches (ETTHs)
(which, in my experience, are usually of a chronic and long-standing nature)
cannot be assessed accurately after only 19 weeks, especially when limited
techniques had been applied 11 weeks earlier. It would be enormously helpful
if future research were to compare different types of manipulation as well
as examine additional (or perhaps more) relevant outcomes. My patients and
I still believe that manipulation has tremendous value in treating ETTH. Let's
not give up.
Michael Felder, DO, MA
Rhode Island Hospital Medical Foundation
Cranston
To the Editor: Several concerns are apparent
after review of the study of spinal manipulation in the treatment of ETTH
by Bove and Nilsson.1
First, in our viewpoint, the "toggle recoil technique" is not an appropriate
manipulative intervention in spinal manipulation of the upper elements of
the cervical spine. Rotary manipulation based on asymmetry of the atlas and
palpable tenderness of the facet joints would be preferable in the treatment
phase of this study. Toggle recoil technique certainly is not a technique
advanced in our curriculum and may provoke pain if excessive force is applied.
A more judicious application of rotary manipulation of the atlas and axis
may have produced a greater therapeutic effect if it was used exclusively
in the study. Furthermore, the low statistical power of this study raises
the possibility of a type II error. The authors assume that a larger study2 demonstrated positive effects of spinal manipulation
due to "personal contact" of the subject and physician. This is opinion and
conjecture on the part of the authors, and such speculation does not invalidate
the results of the larger study. The overall trend in the effects of spinal
manipulation generally is positive in larger, controlled studies.
The authors also mention the possibility that a type II error occurred
due to the small number of participants in the study and that a larger number
of subjects might have identified a treatment effect. Their final statement
that their data suggest that such an effect, if statistically significant,
would be of little clinical significance is a puzzling one.
A 1995 study by Nilsson3 of manipulation
as treatment for cervicogenic headache showed no significant effect of manipulation.
In a 1997 study, Nilsson and colleagues4
observed that the 1995 result could have been due to a type II error (eg,
too few participants). Their 1997 study was then conducted with a larger sample
size, for the express purpose of avoiding a type II error, and a statistically
significant effect of manipulation for cervicogenic headache was demonstrated.
Rod L. Kaufman, DC
Paul Delaney, DC, PhD
Glendale Chiropractic Clinic
Glendale, Calif
In Reply: We thank the authors of these letters,
since as practitioners they represent the people our findings affect. However,
the letters indicate overinterpretation and misinterpretation of our results,
something we avoided. Perhaps more important, the authors collectively fail
to acknowledge that we have years of experience treating headaches using various
manual therapies. In our practices, our results and discussions with others
indicated ambiguity regarding the response of headaches to manual therapy,
and we both became scientists to look for answers. As chiropractors, we would
have preferred a positive outcome, of course. As scientists, we just looked
for the truth.
Space does not permit us to answer all the concerns raised in these
3 letters. Any discussion of reflexogenic pathways and altered biomechanical
integrity should be saved for the appropriate theoretical venue. However,
some issues warrant attention because they indicate prevalent misunderstandings.
The manipulative techniques used in this study reflected the majority
of practices and teaching methods. Our practitioner used a variety of techniques,
indicated by the "gold standard" of palpation and applied using years of accumulated
experience. It is unlikely that the choice matters, however, since most therapists
use strikingly similar techniques, especially in the cervical spine. Moreover,
there is no indication that one technique that gaps the joint is superior
to another, although most chiropractors would agree that gapping the joint
is important. We remind readers that any adjustment, applied injudiciously,
can provoke pain and other symptoms.
The possibility of a type II error always exists; we minimized ours
using a statistical power of 90% to detect a 1-hour difference in daily headache
hours. But our point here, really, is that even if a statistically significant
treatment effect is found in a randomized controlled trial, that treatment
effect may not be large enough to be of practical clinical interest (ie, "of
clinical significance"). For example, in our research, a larger sample size
may have shown a statistically significantly reduced analgesic intake. However,
even at the extremes of our confidence intervals, this would not have amounted
to more than a half dose of the analgesics.
Headache diagnosis is fraught with pitfalls, and tension and cervicogenic
headache presentations often overlap. In our experience, many field practitioners
are not aware of the distinctions between these specific diagnostic categories,
yielding statements like " . . . ETTHs resulting from cervical joint dysfunction."
Headaches resulting from cervical joint dysfunction are cervicogenic headaches. This is important because this group of patients
responds remarkably well to manipulation. Our combined clinical and scientific
experience suggests that if a manipulation treatment helped a tension-type
headache, it probably was a misdiagnosed cervicogenic headache.1
We suggest that practitioners learn to differentiate cervicogenic and
tension-type headaches, since they clearly demand different treatments, and
the 2 types together amount to about 80% of all headache cases.2, 3
We also suggest that practitioners apply the information contained in research
articles judiciously and without emotion, and consider, when possible, the
intent of the authors.
Geoffrey Bove, DC, PhD
Beth Israel Deaconess Medical Center
Boston, Mass
Niels Nilsson, DC, MD, PhD
Odense University
Odense, Denmark
