To The Editor:
"Failure of External Spinal Skeletal Fixation to Improve Predictability of Lumbar Arthrodesis" (2001:83:1656-9), by Bednar, represents a long experience with use of external spinal skeletal fixation as a tool to select candidates for lumbar arthrodesis for the relief of back pain. We strongly disagree with the author's general conclusion that external spinal skeletal fixation has poor predictive value for relief of back pain after spinal arthrodesis. Dr. Bednar pooled data from five clinical studies on this technique 1-5 to support his conclusion. From an epidemiological point of view, this method of data extraction is not correct. When external spinal skeletal fixation is used as a test for the selection of suitable candidates for arthrodesis, the element of a nonfixation control is extremely important. Studies that include use of the nonfixation position, as defined by Esses et al. 1 , in this test must be considered.
In our study, published in 1999 5 , the issue of nonfixation was addressed. Pain relief after successful fusion was obtained in thirty-five (64%) of fifty-five patients. After they knew the final result, forty-three patients (78%) reported that they would have the same procedure, including the temporary external spinal skeletal fixation test, again.
In our study, 67% of the fifty-five patients who went on to fusion had been considered failed back-surgery cases. We believe that, for this group of patients, the fixation test was extremely helpful in our selection of the appropriate candidates for arthrodesis.
The study by Dr. Bednar is definitely flawed given its failure to include the control unlocked position of the fixator in the external spinal skeletal fixation test. This may explain to some extent the low success rate that he reported in this paper.
D.A. Bednar replies:
Drs. Pavlov, de Kleuver, Spruit, and van der Schaaf may, in fact, be right. However, I don't think so.
It can be argued that my study was flawed by the absence of a placebo or "unlocked" trial of the percutaneous external lumbar fixation device. This concern has been raised, but not consistently, by previous authors of earlier published protocols.
I did not include an unlocked control test in my protocol for two reasons. The first is the ethical issue of submitting a patient to an invasive procedure with potential complications only to provide a study control; I did not feel that would be appropriate. The second was, of course, the more obvious question of how, in fact, one could blind a patient to the unlocked control position of the fixator. I can conceive of no way that this could be done without its being obvious to the patient, and, accordingly, I would expect that lack of objectivity regarding the "control" procedure to render the results of any such control unreliable.
At any rate, it probably does not make a difference.
Dr. Pavlov points out that the results of his study, even with the benefit of the "control" unlocked test, resulted in clinically successful pain relief in only 63% of his study population 5 . That result is little different from the results of my study. Frankly, I think that neither of our results are good enough, and, accordingly, I stand by my earlier conclusion that external spinal skeletal fixation is probably not an acceptable clinical tool at this time.