To The Editor:
The article "Flexible Intramedullary Nailing for the
Treatment of Unicameral Bone Cysts in Long Bones" (82-A: 1447-53,
Oct. 2000), by Roposch et al., brings to the forefront an important issue
in pediatric orthopaedics—the necessity of evidence-based
medical care. Roposch et al. reported on the use of flexible intramedullary
nailing to stabilize pathological fractures and to treat cysts without
current fractures. Although the results of this new treatment are
encouraging (a 94% healing rate within thirty-six months),
they are reminiscent of the initial reports on bone marrow injection.
Lokiec et al.1 initially reported
healing rates of 100% with one bone marrow injection, but clinical
practice and a subsequent report would suggest lower healing rates
of 67%2. Articles such
as the one by Roposch et al. exert a persuasive influence on the clinical
community and often purvey the "treatment du jour."
Many theories have been proposed to explain the pathogenesis
of simple bone cysts. The uncertainty of the etiology is reflected
in the numerous treatments presented in the past decade: high-porosity
hydroxyapatite cubes, trepanation, Ethibloc injections, steroid
injection, cryosurgery, bone marrow injection, demineralized bone
matrix, intramedullary nailing, and the various combinations of
multiple treatments or experimental models not yet published.
If we are to bring evidence-based medical care to our practice
of pediatric orthopaedics, we must perform randomized clinical trials.
Furthermore, many of the conditions treated by pediatric orthopaedic
surgeons, such as simple bone cysts, are relatively rare. Therefore,
it will be necessary to perform multicenter studies in order to accrue
a sufficiently large cohort of patients. The POSNA (Pediatric Orthopaedic
Society of North America) Clinical Trials Network represents a society-wide
effort to perform such trials. For trials to be successful, surgeons must
hold off on adopting the "treatment du jour" and
evaluate the current treatments, such as those for simple bone cysts,
in randomized clinical trials. The purpose of clinical trials is
not to stifle scientific creativity but to evaluate the effectiveness
of treatment in standard clinical practice and to provide rapid
and definitive resolutions to controversial clinical dilemmas.
A. Roposch, V. Saraph, and W.E. Linhart reply:
We thank Ms. Donaldson and Dr. Wright for their comments. The
study that we have presented is a case-control study in which the
outcome of a treatment modality for the management of unicameral
bone cysts is evaluated. Treating unicameral bone cysts with flexible
intramedullary nailing permits decompression of the cyst, provides early
stability to the bone, permits early mobilization, decreases the
need for splints/casting for long periods, and decreases
the incidence of the most common complication—a pathological fracture.
In addition, this method of treatment involves minimal surgical intervention,
which is well tolerated by children.
Since this is a case-control study, comparison with the results
of other treatment modalities has not been made. Future studies
comparing the outcome of intramedullary nailing with those of different
treatment methods are needed. We agree with Ms. Donaldson and Dr. Wright
that randomized trials are necessary to establish the effectiveness
of a particular treatment. However, we believe that case-control
studies also have valid clinical implications. They provide a way
of presenting results of new treatment techniques and form a foundation
for future randomized, controlled studies. Every study design has its
limitations and its advantages. In a case-control study like ours,
there is the possibility of selection bias. Multicenter studies
have the advantage of collecting data on more subjects for evaluation,
particularly for rare conditions such as bone cysts. However, multicenter
studies also have their limitations. When data from different surgeons
and different institutions are evaluated, factors such as postoperative management
and supplementary treatment, which may be quite variable, would
also influence the outcomes.
We wish to congratulate Ms. Donaldson and Dr. Wright for their
efforts to perform multicenter studies through the POSNA Clinical
Trials Network, and we would definitely be interested in contributing
our efforts toward the practice of evidence-based medicine. We are
of the opinion that contacting university hospitals directly would
be a more effective way to bring different institutions to collaborate
in establishing evidence-based treatment regimens.