To The Editor:
I would like to express my appreciation and admiration of "Percutaneous
Screw Fixation or Cast Immobilization for Nondisplaced Scaphoid
Fractures" (2001;83:483-8), by Bond et al. The study, which
was very carefully designed, raises important issues. It also illustrates
some of the difficulties of this type of investigation and of the study
of scaphoid fractures in particular.
Randomized trials comparing operative and nonoperative treatment
are prohibitively difficult to perform because patients are unwilling
to leave such an important treatment decision to chance. The authors
were able to enroll only twenty-five of sixty-two eligible patients
under what might be considered ideal circumstances—a captive military
population. I have been much less successful with such trials in
my hospital, and I am grateful to the authors for their achievement.
Alternative study methods such as open-label trials with careful
accounting for important variables may be necessary to compare operative
and nonoperative treatment.
The distinction between displaced and nondisplaced fractures
may be paramount, and the absence of such a distinction limits the
usefulness of most prior studies of scaphoid fractures. Unfortunately,
plain radiographic criteria may not be adequate to identify displacement.
For instance, the fracture in Figure 1 is displaced by at least
1 mm, and, without use of computed tomography, it is not possible
to tell how much angulation and displacement are present at the
fracture site.
Fracture union was used both as a criterion for cast removal
and as an outcome variable. Time to union is an imprecise outcome
measure, particularly when evaluating the scaphoid. Several studies have
questioned the reliability of interpretations of union derived from
plain radiographs. Time to union is probably even less reliable
as an outcome measure after operative treatment.
Return to work is also a difficult outcome measure to interpret.
The physician directly controls return to work through decisions
regarding how to interpret union and whether or not to allow patients
to work with the arm in a cast or splint. I would argue that the earlier
return to work in this study was predetermined and a foregone conclusion,
since the purpose of screw insertion in a nondisplaced fracture
is to limit the time that the patient spends with the arm in a cast
or splint and to allow an earlier return to work or sports.
The data on motion and grip strength are welcome since most prior
studies have focused primarily on union as an outcome measure. What
this study tells me is that if you are sure that an isolated scaphoid
fracture is nondisplaced, it will heal if protected adequately,
and that very little motion or grip strength will be lost in spite
of prolonged cast wear. It probably does not matter much whether
or not the thumb or the elbow is included in the cast.
Percutaneous screw fixation of the scaphoid has been simplified
by the development of small-diameter cannulated screws. Surgeons
should realize that use of these screws is fraught with peril. They
utilize a small-diameter (approximately 1-mm) wire that can be bent
as the arm is moved repeatedly during checking of the position of
the wire. Drills used for some screws are large and clumsy and can
very easily shear a bent wire if forced over it, while drills for
other screws are very narrow and delicate and can explode if so forced.
To avoid problems, great care must be taken in performing this apparently
simple procedure.
I believe that percutaneous screw fixation is an excellent option
for certain patients who want to avoid prolonged cast wear and/or
want to return to certain sports (such as basketball) sooner. These
patients must be willing to accept the risks related to the operative
procedure.
C.D. Bond, A.Y. Shin, M.T. McBride, and K.D. Dao reply:
We thank Dr. Ring for his analysis and comments. Although we
agree with Dr. Ring that randomized trials comparing operative and
nonoperative treatment of scaphoid fractures are difficult to perform,
we disagree with his statement that our patient population was a "captive
military population." Each active-duty military patient
who met the inclusion criteria for the study received an informed-consent
form and was able to choose whether or not he or she would participate.
Even after enrolling in the study, patients were free to withdraw without
penalty. The active-duty military population is an ideal cohort
for randomized studies as it comprises a homogeneous group of individuals
with guaranteed medical care and essentially standard criteria for
return to work worldwide. These patients are motivated to return
to work, as "limited duty" status removes the
patient from the workplace and often dictates a move into temporary
quarters, with a predictably negative effect on morale. So, this
military population is an ideal study population, but far from "captive."
We agree that determining scaphoid fracture displacement can
be difficult; however, in our determination of fracture displacement,
we obtained five radiographic views of the scaphoid (anteroposterior,
lateral, ulnar deviation, radial deviation, and clenched-fist) at
each follow-up visit and applied the strictest radiographic criteria
for scaphoid waist fractures1.
In Figure 1-A, an anteroposterior radiograph with the wrist in ulnar
deviation showed a fracture that was distracted approximately 1
mm, but there was no displacement or "step-off" shown
in this figure or in the others, which clearly demonstrated the
radiographic outline of the scaphoid without displacement. Although
it has been suggested that computed tomographic scanning may be
necessary to determine if a fracture is nondisplaced2, there are no published studies that compare
a five-view radiographic series with planar computed tomographic scans
of nondisplaced scaphoid fractures. To our knowledge, the published literature
on the use of computed tomographic scans for assessment of scaphoid
fractures addresses displaced fractures, nonunions, malunions, and healing
criteria, but not nondisplaced fractures1,3,4.
We agree that fracture union and return to work are both difficult
criteria to measure, as we stated in our paper. Potential observer
bias may exist in the evaluation of the radiographs, as the radiographs
could not be blinded with respect to surgical treatment versus cast immobilization.
However, the criteria for fracture union (a nontender fracture site
in addition to bridging trabeculae shown on the five-view radiographic series
of the scaphoid) were strict and paralleled the criteria used for
fracture union in other published reports of fracture healing5-7. The criteria for return to work
were the same in all patients: a healed fracture and the ability
of the patient to perform without modification the job held prior
to the injury. The military setting is rigid and does not tolerate
the modification of duties without medical clearance. Thus, despite
possible surgeon bias, the military environment dictated the time
of return to work. Finally, this was designed as a prospective and
randomized study to compare short and long-term outcomes of percutaneous cannulated
screw fixation with those of cast immobilization of stable scaphoid fractures.
The conclusions were not foregone or predetermined, as Dr. Ring suggests.
Also, no conclusions regarding the inclusion or exclusion of the thumb
or elbow in the cast can be made from our data.
We are pleased that Dr. Ring agrees with our assertion that this
technique is not without risk. The technique has a learning curve
that is unforgiving if performed improperly or cavalierly. We do not
recommend surgical fixation for all stable scaphoid waist fractures.
Our current practice is to perform the procedure only on patients
who require a rapid or time-sensitive return to work or athletics
and who fully understand the potential complications of this surgical intervention.