To The Editor:
We read with great interest the article entitled "Operative Compared
with Nonoperative Treatment of a Thoracolumbar Burst Fracture without
Neurological Deficit. A Prospective, Randomized Study" (2003;
85:773-81), by Wood et al. The prospective, randomized design of this study
sets it apart from the numerous retrospective case series that have been
published in the last two decades and is generally regarded as the only way to
obtain valid results1-3. Since the debate about the operative or
nonoperative treatment of traumatic burst fractures has not been resolved
because of weak scientific evidence from both sides, the study by Wood et al.
could have provided us with some solid claims that could have guided us to an
optimal treatment regimen. Therefore, we were a bit disappointed to find that
the study had some serious flaws that, in our opinion, render the authors'
conclusions unjustified.
First, the inclusion and exclusion criteria were not described in
sufficient detail. The study included patients with an isolated "burst
fracture," which, in our opinion, comprises a wide range of fracture
types with varying severity. Since fracture classification schemes have been
designed and validated to help one to choose the appropriate treatment and to
predict clinical outcome, their use should be encouraged
strongly4-6. Furthermore, since the authors used plain radiography
and computed tomography but not magnetic resonance imaging to classify the
fracture, the correct exclusion of patients with "a loss of structural
integrity within the posterior osteoligamentous complex" can be in
doubt7. Tables E3 and E4 in the appendix of the study list two
patients (Case 8 from the operative treatment group and Case 8 from the
nonoperative treatment group, both of whom had a fracture at L1) who had
kyphosis angles of 32° and 30°, respectively, at the time of
admission. It is physically impossible to have a kyphosis angle of this degree
with an intact posterior ligament complex at this level. Additionally, eleven
patients from the operative group and seven patients from the nonoperative
group had kyphosis angles of >15°, and some of these patients could
have had AO type-B fractures that were not recognized on radiographs or
computed tomographic scans.
Injuries other than the spinal fractures that were the direct result of the
trauma were not mentioned, although their presence could have seriously
affected the clinical outcome and the responses on questionnaires at the time
of admission and during follow-up.
Patients could be enrolled in the study up to three weeks after the time of
injury. Did not these patients receive treatment during this period? Did these
patients convert to a different treatment regimen after randomization? If a
serious medical condition such as an unstable spinal fracture is diagnosed,
treatment (whether operative or nonoperative) should start as soon as
possible. A surgical reduction performed three weeks after trauma is much less
likely to be successful than one performed on the day of trauma.
Unfortunately, the authors did not provide us with the answers to these
questions.
The patients indicated their perception of pain on a visual analog scale at
the time of admission to the trial. Did the patients receive analgesics after
enrollment or before? Again, we are left in doubt as to whether the fracture
was fresh or days to weeks old, which could have affected the patients'
perception of pain. The participants were initially managed with bed rest for
two to five days after enrollment. If taken to the extreme, it was therefore
possible that patients underwent surgery up to twenty-six days after the
initial trauma. Depending on randomization and the surgeon's preference, the
patients were treated with one of several options: a body cast, a
thoracolumbar orthosis, short-segment (two discs bridged) pedicle-screw hook
instrumentation, long-segment (more than two discs bridged) pedicle-screw hook
instrumentation (mistakenly referred to as a short-segment technique), or an
anterior fibular and rib-strut construct arthrodesis. This multitude of
treatment options does nothing to clarify the optimal treatment regimen. Why
did a surgeon choose a demanding anterior approach if "no formal attempt
was made to decompress the neural canal," which is its foremost
advantage over posterior instrumentation? Since pedicle-screw systems and
reports describing their use, complications, and outcome were widely available
by 1994 when the study was started, we do not comprehend completely why more
than one surgical option was allowed, resulting in a very heterogeneous
population8-21. If Figures 2-D and 2-E are exemplary of the
authors' surgical approach, it is easy to understand why some patients in the
operatively treated group had low-back pain as laminar hooks in the mobile
lumbar spine are known causes of discomfort and pain. If one wants to compare
two treatment modalities, one should use the contemporary state-of-the-art
techniques of both treatment options. There is a substantial difference in the
clinical results of surgical treatment of traumatic spine fractures between
the current series and other series22-26. We do not think that the
results for the operative treatment group in this study represent the outcomes
that have generally been reported in association with short-segment fusion
techniques. The authors should clarify these differences. Nineteen
complications were reported for the operatively treated group, and two were
reported for the nonoperatively treated group. It is our opinion that,
technically speaking, complications are medical conditions that occur and can
be diagnosed in the postoperative period (the length of this period can be
discussed, but most physicians would probably say that it falls somewhere
between one and three months). Instrumentation/bone failure or instrumentation
breakage is therefore not a complication, whereas a resulting kyphosis, for
example, is. Similarly, the removal of instrumentation is not a complication;
it is part of the surgical treatment or the result of an infection or other
medical condition. Since these medical conditions were not mentioned, the
complication count is reduced to nine. The authors conclusively state that
operative treatment did not provide an advantage over nonoperative treatment.
In light of the above concerns and the limitations discussed by the authors,
we are not convinced that this claim can be justified. The authors' attempt,
praiseworthy though it is, to provide us with solid information regarding the
optimal treatment regimen for burst fractures without neurological deficits
has therefore missed its goal.
We thank Drs. Verlaan and Oner for their comments regarding our study. We
will address the comments as they appeared in their letter.
Drs. Verlaan and Oner suggest that our inclusion and exclusion criteria
were not described in sufficient detail. Our reply is that we believed that we
provided ample criteria both clinically and radiographically to define our
study population. The text of the article indicates that the patients had
stable burst fractures without posterior ligamentous disruption, neurologic
sequela, or posterior osseous fracture or dislocation, which is consistent
with the AO type-A3 fracture classification. We use the AO system in our work.
The article by McCormack et
al.4 that was cited
by Drs. Verlaan and Oner is not a classification system but is a useful tool
for planning surgery, and the article by Oner et
al.6 is a magnetic
resonance imaging review of thoracolumbar fractures without any reference to
patient-related outcome. Finally, we felt comfortable reporting that these
individuals did not have any substantial loss of structural integrity on the
basis of the lack of any plain radiographic or computerized tomographic
findings suggestive of facet disruption or interspinous widening as well as a
careful clinical examination. The fact that two patients had kyphosis angles
of 32° and 30° at the time of admission is not physically impossible.
Within this patient population, kyphosis angles may reflect the morphology of
the adjacent vertebrae, whether altered or not. The eleven patients in the
operative group and the seven patients in the nonoperative group with kyphosis
angles of >15° could have had type-B fractures that were unrecognized,
but none went on to have any of the sequelae of posttraumatic kyphosis that
are consistent with unrecognized ligamentous injuries and we stand by our
examinations.
Drs. Verlaan and Oner describe the possibility that injuries other than the
spinal fractures could have influenced the long-term clinical outcome as
reflected in questionnaires. This may be true, especially for questionnaires
such as the SF-36. However, the injuries were equally divided between each
group and this bias essentially cancels itself out. The other questionnaires
were aimed specifically at back pain and back disability.
Patients were allowed to be enrolled in the study up to three weeks after
the time of the injury. In actuality, all were enrolled within one to three
days after the injury; therefore, none of the listed concerns are pertinent to
this study.
In response to the question regarding the various treatment options (i.e.,
anterior versus posterior), we would suggest that three disc spaces being
bridged does not constitute a long-segment pedicle screw instrumentation as
opposed to two disc spaces being crossed. Our original intention was simply to
answer the question of whether operative treatment provided different outcomes
than did nonoperative treatment with a cast or brace, not whether operative
treatment (or which operative technique) is preferable; that subject has been
addressed in a follow-up study pending publication. We do not believe that an
anterior approach is "demanding" since the time in the operating
room and the rates of transfusions and complications associated with the
anterior approach are no different from those associated with the posterior
approach. We agree that it would have been theoretically a somewhat cleaner
study if only one operative technique had been used, but, as stated above,
there is very little difference between the two approaches when properly
performed.
Drs. Verlaan and Oner suggest that laminar hooks in the mobile lumbar spine
are known causes of discomfort and pain. We are aware of literature suggesting
that laminar hooks placed into the caudad aspect of the lumbosacral area have
a higher risk of mechanical failure; however, we are not aware of any
publication that has specifically addressed placement of a laminar hook into
the lumbar spine as a source of postoperative discomfort and pain. All of the
hooks in our study were at L3 at the most caudad point and did not involve the
more caudad aspect of the lumbar spine. We have come to believe that, in many
instances, chronic lumbar pain following surgical treatment of thoracolumbar
fractures may be related more to preexisting degeneration of the lumbosacral
disc space than to the type of instrumentation or instrumentation technique
used at the thoracolumbar junction.
There has been no obvious consensus as to what constitutes the
postoperative period. We disagree with the assertion that most physicians
would say that this period falls somewhere between one and three months after
the procedure. Late infection following deformity surgery is certainly a
complication and can take place years after the procedure. We therefore
believe that instrumentation/bone failure or instrumentation breakage is a
complication, especially if it is associated with symptoms. Similarly, if
postoperative instrumentation is thought to be a source of pain requiring
surgical treatment, we believe that it is a complication. This latter point,
however, is certainly open for interpretation. It does make for another
hospital stay, more morbidity, more cost, and more risk, and therefore we took
the liberty of calling it a complication.
With regard to the final point about our conclusions, we believe that if
one were to remove the concerns (imagined or not) regarding our
inclusion/exclusion criteria, complication list, and so on, the fact remains
that the function scores, pain scores, long-term outcome, SF-36 scores, and
return to work are all so similar that we could not conclusively state that
operative treatment provided any distinct advantage over nonoperative
treatment.
Bhandari M, Morrow F, Kulkarm AV,
Tornetta P 3rd. Meta-analyses in orthopaedic surgery. A systematic review
of their methodologies. J Bone Joint Surg Am.2001;83:
15-24.8315
2001
[PubMed]
Kiter E, Karatosun V, Gunal I. Do
orthopaedic journals provide high-quality evidence for clinical practice?
Arch Orthop Trauma Surg.2003;
123: 82-5.12382
2003
[PubMed]
Concato J, Shah N, Horwitz RI.
Randomized, controlled trials, observational studies, and the hierarchy of
research designs. N Engl J Med.2000;342:
1887-92.3421887
2000
[PubMed][CrossRef]
McCormack T, Karaikovic E, Gaines
RW. The load sharing classification of spine fractures.
Spine.1994;19:
1741-4.191741
1994
[PubMed][CrossRef]
Magerl F, Aebi M, Gertzbein SD, Harms
J, Nazarian S. A comprehensive classification of thoracic and lumbar
injuries. Eur Spine J.1994;3:
184-201.3184
1994
[PubMed][CrossRef]
Oner FC, van Gils AP, Dhert WJ,
Verbout AJ. MRI findings of thoracolumbar spine fractures: a
categorisation based on MRI examinations of 100 fractures. Skeletal
Radiol.1999;28:
433-43.28433
1999
[CrossRef]
Leferink VJ, Veldhuis EF, Zimmerman
KW, ten Vergert EM, ten Duis HJ. Classificational problems in ligamentary
distraction type vertebral fractures: 30% of all B-type fractures are
initially unrecognised. Eur Spine J.2002;11:
246-50.11246
2002
[PubMed][CrossRef]
Aebi M, Etter C, Kehl T, Thalgott
J. Stabilization of the lower thoracic and lumbar spine with the internal
spinal skeletal fixation system. Indications, techniques, and first results of
treatment. Spine.1987;12:
544-51.12544
1987
[PubMed][CrossRef]
Akalm S, Kis M, Berth IT, Citak M,
Mumcu EF, Tuzuner M. Results of the AO spinal internal fixator in the
surgical treatment of thoracolumbar burst fractures. Eur Spine
J.1994;3:
102-6.3102
1994
[CrossRef]
Bednar DA. Experience with the
"fixateur interne": initial clinical results. J Spinal
Disord.1992;5:
93-6.593
1992
[CrossRef]
Benson DR, Burkus JK, Montesano PX,
Sutherland TB, McLain RF. Unstable thoracolumbar and lumbar burst
fractures treated with the AO fixateur interne. J Spinal
Disord.1992;5:
335-43.5335
1992
[CrossRef]
Carl AL, Tromanhauser SG, Roger
DJ. Pedicle screw instrumentation for thoracolumbar burst fractures and
fracture-dislocations. Spine.1992;17(8 Suppl):
S317-4.17S317
1992
[PubMed][CrossRef]
Esses SI. The AO spinal internal
fixator. Spine.1989;14:
373-8.14373
1989
[PubMed][CrossRef]
Huler RJ, Esses SI, Botsford DJ.
Work status after posterior fixation of unstable but neurologically intact
burst fractures of thoracolumbar spine. Paraplegia.1991;29:
600-6.29600
1991
[PubMed][CrossRef]
Karlstrom G, Olerud S, Sjostrom
L. Transpedicular fixation of thoracolumbar fractures. Contemp
Orthop.1990;20:
285-300.20285
1990
Kuner EH, Kuner A, Schlickewei W,
Mullaji AB. Ligamentotaxis with an internal spinal fixator for
thoracolumbar fractures. J Bone Joint Surg Br.1994;76:
107-12.76107
1994
[PubMed]
Lindsey RW, Dick W. The fixateur
interne in the reduction and stabilization of thoracolumbar spine fractures in
patients with neurologic deficit. Spine.1991;16(3 Suppl):
S140-5.16S140
1991
[PubMed][CrossRef]
Marti Garin D, Villanueva Leal C,
Bago Granell J. Stabilization of the lower thoracic and lumbar spine with
the internal spinal skeletal fixation system and a cross-linkage system. First
results of treatment. Acta Orthop Belg.1992;58:
36-42.5836
1992
[PubMed]
McNamara MJ, Stephens GC, Spengler
DM. Transpedicular short-segment fusions for treatment of lumbar burst
fractures. J Spinal Disord.1992;5:
183-7.5183
1992
[PubMed][CrossRef]
Sim E, Stergar PM. The fixateur
interne for stabilising fractures of the thoracolumbar and lumbar spine.
Int Orthop.1992;16:
322-9.16322
1992
[PubMed]
Viale GL, Silvestro C, Francaviglia
N, Carta F, Bragazzi R, Bernucci C, Maiello M. Transpedicular
decompression and stabilization of burst fractures of the lumbar spine.
Surg Neurol.1993;40:
104-11.40104
1993
[PubMed][CrossRef]
Parker JW, Lane JR, Karaikovic EE,
Gaines RW. Successful short-segment instrumentation and fusion for
thoracolumbar spine fractures: a consecutive 41/2year series.
Spine.2000,25:
1157-70.251157
2000
[PubMed][CrossRef]
Knop C, Fabian HF, Bastian L, Blauth
M. Late results of thoracolumbar fractures after posterior instrumentation
and transpedicular bone grafting. Spine.2001;26:
88-99.2688
2001
[PubMed][CrossRef]
Sjostrom L, Jacobsson O, Karlstrom G,
Pech P, Rauschning W. Spinal canal remodelling after stabilization of
thoracolumbar burst fractures. Eur Spine J.1994;3:
312-7.3312
1994
[PubMed][CrossRef]
Leferink VJ, Zimmerman KW, Veldhuis
EF, ten Vergert EM, ten Duis HJ. Thoracolumbar spinal fractures:
radiological results of transpedicular fixation combined with transpedicular
cancellous bone graft and posterior fusion in 183 patients. Eur
Spine J.2001;10:
517-23.10517
2001
[CrossRef]
Speth MJ, Oner FC, Kadic MA, de Klerk
LW, Verbout AJ. Recurrent kyphosis after posterior stabilization of
thoracolumbar fractures. 24 cases treated with a Dick internal fixator
followed for 1.5-4 years. Acta Orthop Scand.1995;66:
406-10.66406
1995
[PubMed][CrossRef]