During an eighteen-year period between 1980 and 1998, 473 consecutive
elbow arthroscopies were performed in 449 patients by twelve different
orthopaedic surgeons. The majority (89%) of the arthroscopic procedures,
however, were performed by two experienced elbow surgeons, one of
whom was also fellowship-trained in arthroscopy. Of the other ten
surgeons, two had performed ten to fifteen elbow arthroscopies and
eight had performed fewer than ten. The records of these patients
were retrospectively reviewed to determine the final d iagnoses,
types of procedures, and complications. The surgical records on
all 473 arthroscopic procedures were reviewed to determine the
types of procedures, the type of anesthetic (general, regional,
and/or local), tourniquet time, portals utilized, technical difficulty
(related to visibility during the procedure), and any comorbidities
of the patient. The comorbidities were defined as conditions that
might increase the risk of neuro vascular injury during the arthroscopy
because of altered anatomy. These comorbidities included rheumatoid
arthritis, a subluxating or transposed ulnar nerve, elbow contracture,
prior elbow surgery, and prior elbow trauma. These categories were
not mutually exclusive, and many patients had more than one comorbidity
identified prior to elbow arthroscopy. Each comorbidity was analyzed
separately.
Seventy-eight of the procedures that began with an elbow arthroscopy
included an arthrotomy as well. Of these arthrotomies, forty-six
were performed after a diagnostic arthroscopy and twenty-nine were performed
because the attempt to perform the procedure arthroscopically had
failed. These arthrotomies were performed early in the series and
were most often the result of swelling and poor visibility. The
other three open procedures were done to assist with internal fixation
of an intra-articular fracture of the elbow joint.
In this series, we identified major complications as (1) permanent
nerve injury, (2) compartment syndrome of the arm or forearm, (3)
postoperative joint infection related to the procedure, (4) vascular injury,
and (5) loss of motion exceeding 30°. Minor complications included
(1) transient nerve palsy that completely resolved, (2) prolonged
drainage lasting more than five days or superficial infection at
a portal site, (3) instrument breakage, and (4) loss of motion of
30° or less. Complications were classified as (1) those occurring
during the procedure and identifiable immediately postoperatively
(nerve injury, compartment syndrome, hematoma, and instrument breakage)
and (2) those that developed in a delayed fashion (loss of motion,
joint infection, and persistent drainage or superficial infection
at a portal site).
All 473 procedures were considered in the evaluation of the immediate
complications, whereas only the 414 with a minimum of six weeks
of follow-up were considered in the analysis of the delayed complications.
A minimum of six weeks was considered adequate for the development
of delayed complications. All 473 cases were included in the statistical
analysis of risk factors possibly associated with immediate, operative
neurovascular complications.
An independent observer who did not take part in any of the arthroscopic
procedures (E.W.K.) reviewed all of the patients' charts and analyzed
the data. Patients with less than six weeks of follow-up documented
in the record were contacted by means of a mailed survey. If the
returned survey suggested any complications, then a follow-up telephone interview
was conducted. Of the 473 cases reviewed, 261 had documentation
of a follow-up examination performed after more than six weeks. In
the remaining 212 cases, with less than six weeks of follow-up,
the patient was sent a survey inquiring about any complications.
In 116 cases the patient returned the survey, and in four of these cases
additional telephone follow-up was required to clarify answers.
In thirty-seven of the ninety-six cases in which the survey was
not returned, the patient was successfully contacted by telephone.
In the remaining fifty-nine cases, the patient could not be reached
and was lost to follow-up.
During the telephone interviews, patients were asked about infection,
drainage, subjective weakness, and numbness. If a patient described
a loss of motion in the telephone interview, the amount was estimated
on the basis of the patient's comparison with the contralateral,
or "normal," side, with full extension measuring 0° and the ability
to touch the ipsilateral shoulder considered 130° of flexion. On the
initial review of the patients' histories, it was noted that a loss
of motion had been found at the most recent examination following
fourteen arthroscopic procedures. In seven cases, the patient stated,
during a later telephone interview, that the motion of the elbow
had returned to the preoperative level. In five of the remaining
cases, the contracture persisted at a follow-up physical examination
performed more than six months postoperatively. Two patients with
a contracture at six weeks postoperatively were unavailable for
further examination; they also could not be contacted by telephone,
and they did not return the survey. For the purpose of the analysis,
we assumed that the contracture persisted in these two cases lost
to follow-up.
Of the 473 procedures, 446 (94%) were performed with the patient
under general anesthesia, whereas only twenty-seven (6%) were performed
with a regional block. At the end of 117 (25%) of the procedures
the patient received intra-articula r Marcaine (bupivacaine) and
Celestone (betamethasone), at the end of 135 (29%) of the procedures
the patient received intra-articular Marcaine alone, and at the
end of forty-two (9%) of the procedures the patient received intra-articular
morphine alone. The total tourniquet time averaged sixty-one minutes, with
a range of eleven to 159 minutes. The tourniquet was deflated after
the dressing had been applied.
One hundred and forty-seven (31%) of the procedures were done
in female patients, and 326 (69%) were done in male patients. The
average age was thirty-six years (range, nine to eighty-eight years). Three
hundred and twenty-three procedures (68%) were done on right elbows,
and 150 (32%) were done on left elbows. Comorbidities included rheumatoid
arthritis (seventy-five procedures), contracture of the elbow (seventy-three),
ulnar nerve subluxation or transposition (thirty-nine), prior elbow
surgery (eighty-three), and prior elbow trauma (fifty-one). The
primary underlying diagnoses in the patients with elbow contracture included
osteoarthritis (thirty procedures), rheumatoid arthritis (twenty-three),
osteochondritis dissecans (seven), trauma (four), synovial chondromatosis
(three), loose bodies (two), instability (two), and unknown (two).
The final diagnoses (with some patients having more than one
diagnosis) were osteoarthritis (150 procedures), loose bodies (112),
rheumatoid or inflammatory arthritis (seventy-five), osteophytes and/or
impingement (sixty-two), pain without a diagnosis (forty-eight),
osteochondritis dis secans (forty-three), instability (twenty-eight),
lateral epicondylitis (thirteen), plicae (eleven), fracture of the elbow
(seven), chronic valgus overload (six), synovial chondromatosis
(six), and septic arthritis (five).
Techniques
The general operative techniques utilized have been well described
elsewhere18,21,39,40. Most patients
were placed in the lateral decubitus position with the involved
elbow over a padded bolster. The forearm was allowed to swing free,
and the elbow was flexed 90°. The arm was then prepared in the usual
fashion. The forearm was exsanguinated by elevating the limb. A
soft Elastic Bandage (DePuy, Warsaw, Indiana) was then wrapped around
the hand and forearm to within 10 cm of the olecranon. The tourniquet,
which was used routinely, was inflated to 250 mm/Hg. The Elastic
Bandage was left on until the end of the procedure, to limit the
periarticular swelling to the elbow area. When the bandage and tourniquet
were removed, any accumulated edema rapidly dissipated into the
tissues of the forearm and arm. The joint was next distended with
approximately 15 to 25 ml of sterile normal saline solution through
a needle inserted in the direct midlateral portal (the soft spot).
Several different portals were utilized during the elbow arthroscopies,
with the location and number of portals changing over time. In the
early experience, only one or two portals (usually the direct midlateral
and posterolateral) were used. More recently, however, with the
increasing complexity of the procedures performed, the number of
portals has increased (Fig. 1). In addition, more of an emphasis
has been placed on using the more proximal (proximal anterolateral
and anteromedial) portals. Portal placement is determined by careful palpation
of the underlying osseous structures. We do not rely on skin markings,
as they do not correctly indicate the underlying structures after
swelling occurs. We now use retractors to permit visualization in
the anterior compartment, rather than pressurization to accomplish
joint distension. This reduces the risk of edema and allows more complex
surgical procedures to be performed inside the elbow. Portal placement
has been accomplished with both the outside-in and the inside-out
technique, but over time the outside-in technique has become preferred.
Eight procedures were performed on patients with ulnar neuritis
or neuropathy coexisting with osteoarthritis and contracture. In
these patients, the ulnar nerve was transposed subcutaneously through a
10-cm posteromedial skin incision prior to the arthroscopy. The
medial and posterior portals were placed directly through the open
wound while the nerve was gently retracted.
A previously described system for pressurized irrigation was
routinely used and is recommended40.
The arthroscopic procedures included synovectomy (184 procedures),
d衲idement of joint surfaces or adhesions (180), excision of osteophytes
(164), diagnostic arthroscopy (154), loose-body removal (144), plicae
resection (eleven), washout of an infected joint (five), treatment
of lateral epicondylitis (thirteen), and capsular procedures such
as capsular release, capsulotomy, and capsulectomy (seventy-three).
Contractures were released in a stepwise manner. The first step
involves synovectomy and removal of any soft tissue that may block
motion because of its bulk, such as scar tissue in the olecranon
fossa. The second step involves removal of osteophytes from the
olecranon and coronoid process as well as restoration of the normal
depth and contour of the fossae of the olecranon, coronoid process,
and radial head. If motion is still limited, as it almost always
is, the final stage is capsular release.
In eight of the capsular procedures, the radial nerve (three
procedures) and/or the ulnar nerve (seven) was visualized arthroscopically
from inside the joint to permit safe resection of the capsule while the
nerves were observed, under direct vision, at a safe distance from
the cutting instrument. A full description of the technique of arthroscopic
nerve exploration is beyond the scope of this paper. First the location
of the nerve behind the capsule is determined on the basis of the
knowledge of three-dimensional anatomy of the elbow, and then the capsule
is incised with a wide duckbill basket punch biopsy. The ulnar nerve
is also palpated with a blunt probe through the capsule before the
capsule is opened. The soft tissue behind the capsule is dissected
away to create a plane for safe incision of the capsule. The nerve
is then gently palpated through the opening of the capsule, and
the dissection continues until the fat around the nerve is visualized.
The nerve itself is then visualized and, once it is under direct
vision, the rest of the capsule is incised or excised.
In order to assess whether the perceived increase in the technical
demands of the procedures increased the risk of complications, and
of nerve injuries in particular, we developed a simple, preliminary
system for scoring complexity by assigning a numerical value to
each arthroscopic procedure, ranging from diagnostic arthroscopy
through total capsulectomy and arthroscopic nerve exploration.
Statistical analyses were performed with univariate analysis,
with use of the chi-square test, and the Student t test when appropriate.
Each arthroscopic procedure was counted as one case, with the acknowledgment
that a patient may have had more than one arthroscopic procedure
(such as repeated arthroscopy on the same elbow or arthroscopy on both
elbows). Each case was considered to have its own variables (diagnosis,
procedures, tourniquet time, type of anesthetic, and so on).
The most common immediate minor complications were transient
nerve palsies that affected twelve nerves in ten patients. There
were four superficial radial nerve palsies; five ulnar nerve palsies;
and one posterior interosseous, one anterior interosseous, and one
medial antebrachial cutaneous nerve palsy (Table I). The sensory
and motor symptoms resolved within six weeks after the arthroscopy
in all but one case, in which the symptoms required six months to
resolve. We were able to conduct a follow-up physical examination
to document resolution of the symptoms in all but one case; one
patient was contacted by telephone and reported no residual numbness
or paresthesias.
The only factors that were significantly associated with the
development of transient nerve palsy were an underlying diagnosis
of rheumatoid arthritis (p < 0.001), a contracture (p < 0.05),
and the performance of a capsulectomy or capsular release (p < 0.05).
The average total tourniquet time was slightly longer for the patients
with nerve injury (sixty-eight minutes compared with sixty-one minutes
in the patients without nerve injury), although this difference
was not significant.
Although an anteromedial portal had been placed in all of the
cases in which an ulnar nerve palsy developed, compared with only
159 (34%) of the 468 cases without an ulnar nerve injury (p < 0.05),
we believe that this association was a coincidence (with one exception).
There were no permanent neurovascular injuries, compartment syndromes,
or hematomas identified in our series of 473 arthroscopic procedures.
The most frequent delayed minor complication was prolonged clear
or serous drainage from the portal sites lasting more than five
days after twenty-two procedures (5%). Only the direct midlateral
and anterolateral portals were found to be sites of prolonged drainage.
Of the twenty-two procedures followed by persistent drainage, fourteen
did not involve placement of sutures in the portals, four involved
placement of simple interrupted nylon sutures, and three involved
sub cuticular placement of Vicryl (polyglactin) sutures. One procedure
was followed by drainage from the site of a partial arthrotomy,
between the direct midlateral and posterolateral portal sites. Superficial
infection, as defined by erythema and tenderness around the portal
sites, developed after eleven procedures (2%). All eleven of these
infections resolved with a short course of oral antibiotics.
Elbow joint infection developed after four (0.8%) of the arthroscopic
procedures. Prophylactic intravenous antibiotics were given before
only one of these four procedures, whereas they were given prior
to 277 (59%) of the 469 arthroscopies that were not followed by
joint infection (p = 0.17). In addition, joint infection was more
common (four of 117 procedures) when the patient received intra-articular
steroids at the end of the procedure than when steroids were not
used (zero of 356 procedures) (p < 0.005). Fortunately, there
were no permanent sequelae from any of the four joint infections.
An examination was performed at more than one week after 445
of the 473 arthroscopies and at one month or more after 283 of the
operations. Although a loss of motion was demonstrated at the initial
postoperative examination in fourteen of the 445 cases, the loss
was found to have persisted at a later examination or on telephone
interview in only seven cases. In all seven cases with persistent
contracture, there was a minor decrease (5° to 15°) in the range
of motion. Two cases were lost to follow-up after six weeks, and
motion had not been regained by that time. The motion was lost in
the flexion-extension plane in all seven cases. Heterotopic ossification
did not develop in this series, although in two cases a small amount
of bone formation occurred at the tip of the olecranon following
removal of olecranon osteophytes.
During the years of this study, the complexity and extent of
the arthroscopic procedures increased. The average complexity increased
exponentially from 1980 to 1998, with the greatest increases in the
last six years (r2 = 0.96; p < 0.0001;
Fig. 2). From 1980 to 1992, less than three portals were used on
average (Fig. 1). However, from 1993 to 1998, the average increased
from three to five. While there was an increase in the average complexity
of the procedures performed during the eighteen years of this study,
there was a much smaller increase in the total rate of complications
(r2 = 0.13; p < 0.05; Fig. 3). More
importantly, the rate of neurological complications did not increase
as the complexity of the procedures increased (r2 =
0.0023; p = 0.487; Fig. 4). In fact, the rate of nerve palsies declined
slightly.
Elbow arthroscopy is a technically challenging procedure, and
attention to the neurovascular anatomy about the elbow is essential
for the prevention of complications. In contrast to the complication
rates associated with arthroscopy of the knee and shoulder, the
prevalence of complications associated with elbow arthroscopy had
not previously been well defined. Data from surveys may not provide
a reliable indication of either the true frequency of the procedure
or the true prevalence of complications. The Arthroscopy Association
of North America conducted two surveys; one was a retrospective review
of 395,566 arthroscopic procedures41,
and the other was a prospective survey of twenty-one experienced
arthroscopists42. The overall
complication rates for all joints were 0.56% and 1.68%, respectively.
Of the 1648 elbow arthroscopies in these two surveys, only one was reportedly
followed by nerve injury.
The largest single series providing outcome data on elbow arthroscopy
that we found in the peer-reviewed literature was that reported
by two of us (S.W.O'D. and B.F.M.) in 199221.
In that report of seventy-one elbow arthroscopies with a detailed
analysis of untoward events, the overall rate of major and minor
complications was 10%. Schneider et al. reported a similar complication
rate, noting seven temporary nerve injuries in their study of sixty-seven
patients36. Schneider et al. were
concerned about this high rate of nerve injury, stating that "in
the future, the increased use of elbow arthroscopies by inexperienced
surgeons will probably lead to an increase in complications."
In 1996, a review of the literature by Savoie revealed sixteen
reported complications after 465 elbow arthroscopies (a 3% prevalence)43. Savoie noted, however, that all
sixteen complications occurred after procedures performed by eight authors
who had carried out just 127 of the 465 operations reported. Savoie
also reported his own experience with elbow arthroscopy, describing twenty-two
complications after 247 procedures (a 9% prevalence). On the basis
of these results and those described by other authors, the rate
of permanent and transient complications after elbow arthroscopy
appears to be much higher (in the range of 10%21,36,43)
than that reported after arthroscopy of the knee and shoulder (1%
to 2%41,42,44).
In our current series of 473 elbow arthroscopies, we recorded
four types of minor complications in fifty cases (11%). These included
prolonged drainage from the portal sites, superficial infection,
contracture, and temporary nerve palsy. The most frequent minor
complication was prolonged drainage or erythema around the lateral
portal sites, usually when Steri-Strips (3M, St. Paul, Minnesota)
rather than sutures had been used to close the portal sites. In
all of the patients, the drainage subsided after a two to four-week
course of prophylactic oral antibiotics. Persistent drainage from
portal sites was not described as a complication in the reports
on arthroscopy-related complications by the Arthroscopy Association
of North America41,42,44. The
lateral portal sites appear to be the most susceptible to this complication
because the subcutaneous tissues are thin and are unable to provide
a barrier or seal between the distended joint and the skin. We now
routinely use, and recommend the use of, a locked horizontal mattress
stitch for the closure of portal sites to prevent prolonged drainage.
The most serious postoperative complication in the present series
was deep infection, which occurred in four cases (0.8%). This complication
occurred only in patients who received intra-articular steroids
at the end of the procedure. Because of this, we rarely inject corticosteroids
into the joint at the end of our arthroscopic procedures, and our
current practice involves the routine administration of intravenous
prophylactic antibiotics preoperatively to patients who do not have
a known history of allergic reaction to such antibiotics.
Postoperative contracture is a recognized complication of open
elbow surgery. Elbow motion was lost following seven of the elbow
arthroscopies. In each case, the loss was relatively minor (less
than 20°). One of the patients had undergone an open repair of the
medial collateral ligament following a diagnostic arthroscopy, which
may have contributed to the loss of motion. In the remaining six
cases in which motion was lost postoperatively, the patient had
a condition that predisposed to elbow stiffness (inflammatory arthritis
in two, osteoarthritis in two, an osteochondral defect in one, and
osteochondritis dissecans in one). Four of the arthroscopic procedures
had been relatively minor (diagnostic arthroscopy or loose-body
removal), and the explanation for the loss of motion after those
arthroscopies is unclear.
The prevalence of neurological complications after elbow arthroscopy
has ranged from 0% to 14%1,5,10,21,24,36,42,43,45.
In our series, there were no permanent neurological injuries, but
transient nerve palsy was identified after elbow arthroscopy in
ten patients (2%). The significantly higher risk of nerve injury
resulting from elbow arthroscopy is understandable considering the
close proximity of the nerves to the capsule and the portal sites,
particularly the anterolateral portal site1,32,46-49.
The nerve-to-portal distances increase with joint distension, although
the nerve does not move farther away from the cap sule1,32,48,50. Also, capsular distension
is often not possible in elbows with contracture, as they have a
loss of intracapsular capacity averaging 6 ml51.
Nerve injuries associated with elbow arthroscopy have been reported
to result from compression32,33,
injection of local anesthetic1,18,
or direct trauma29,32,34,38. Only
two of the ten cases of transient nerve palsy in our series could
be attributed to direct trauma from the arthroscopic procedure itself.
One radial nerve was probably compressed by instruments during the
procedure, and one ulnar nerve palsy occurred when the anteromedial
portal was placed too far posteriorly. The temporary nerve palsies
in the other eight cases were attributed to intra-articular injection
of a local anesthetic (two cases), prolonged tourniquet compression
(one), forearm compression from wrapping that was too tight (one),
and delayed-onset nerve dysfunction secondary to the use of an indwelling
catheter and continuous brachial plexus block (four). In these last
four cases (two ulnar and two radial sensory palsies), continuous
passive motion was employed in the immediate post operative period.
We believe that these complications may have been due to compression
of the radial nerve or stretching of the ulnar nerve in the continuous-passive-motion device
while the limbs were relatively anesthetized. The two patients with
temporary ulnar neuropathy had greatly increased flexion from the
procedure, which likely permitted stretch-induc ed neuropathy. (We
have seen this same complication after open release.) When this
is anticipated, we now transpose the ulnar nerve subcutaneously.
We no longer inject a local anesthetic following arthroscopy of the
elbow because one patient had a second operation that clearly could
have been avoided had local anesthesia not been used. We did not
identify any cutaneous neuromas, which may be due to the technique
that we use for establishing the portals. Cutting only through the
skin, with the surgeon making certain to drag the skin across the
knife blade rather than making a stab incision, can protect the
superficial nerves from the blade.
For an anteromedial portal we use the proximal anteromedial entry
site, rather than the standard anteromedial portal, to increase
the nerve-to-portal distance and to minimize the amount of manipulation
through the bulky muscles of the flexor pronator origin. We now
reserve the use of the originally described, more distal anteromedial
portal for elbows in which a second retractor is required in the
anterior compartment (with one already placed in the lateral side).
Entry into a contracted joint is best accomplished with use of a
custom-made switching stick that has been machined from a Steinmann
pin to a taper-point at the end. The point is blunt enough so that
it will not cut into tissues, yet tapered enough so that it can
be used to penetrate the capsule without deflecting off of it. Once this
blunted Steinmann pin has been placed, the arthroscope sheath is
slid into the joint, over the pin, and the pin is withdrawn. We
believe this to be easier, safer, and more effective than trying
to place the arthroscope sheath containing an obturator into the
joint.
It is likely that nerve injuries are related more to the use,
than the selection, of instruments. In general, we used a 4.0-mm
arthroscope and motorized shavers ranging from 3.5 to 5.5 mm. Suction
was avoided near the nerves, and motorized instruments were directed
away from the nerves rather than towards them. On two occasions,
while a burr was being used on the medial aspect of the olecranon
in an elbow with advanced osteoarthritis, the capsular tissues were
caught in the burr and drawn into it, which brought the ulnar nerve
dangerously close to the burr. Neither patient had a nerve injury
as a result of the mishap, but the possibility for devastating nerve
injury was apparent.
In our series, two factors that were associated with a higher
risk of nerve palsy were the performance of capsular release and
a diagnosis of rheumatoid arthritis. Capsular release is one of
the therapeutic indications for which the use of elbow arthroscopy is
rapidly increasing. In the present study, four of the ten patients
in whom transient nerve palsy developed had a prior elbow contracture.
In all four of these patients, however, the temporary nerve palsy
was attributed to a mechanism other than direct trauma (intra-articular
injection of anesthetic in one and placement of an axillary catheter
in three). Despite the lack of a direct relationship between nerve
injury and elbow contracture in our series, we believe that a loss
of capsular space increases the difficulty of the procedure. Kim
et al. reported two transient median-nerve palsies after arthroscopic
capsulec tomy in patients with contrac ture10.
Jones and Savoie reported a posterior interosseous nerve transection
in a patient with an elbow contracture who had a capsulectomy performed9. Haapaniemi et al. reported a case
of complete transection of the median and radial nerves in a patient
with posttraumatic elbow contracture treated by arthroscopic capsular
release52.
A review of the literature revealed reports of at least two complete
nerve transections that occurred during synovectomy in patients
with rheumatoid arthritis, who are believed to be at increased risk
for nerve injury during elbow arthroscopy. Ruch and Poehling noted
that patients with rheumatoid arthritis have a "thin and filmy capsule."34 We agree and would add that the
altered osseous architecture impairs the surgeon's ability to predict the
nerve location because of the loss of normal intra-articular landmarks.
In our series, seven of the ten patients with nerve complications
had rheumatoid arthritis, although only seventy-five of the 473 procedures
were in patients with a diagnosis of rheumatoid arthritis. However,
no patient sustained a complete nerve injury or evidence of a nerve
laceration in our series. It has become our routine practice to
release the capsule in rheumatoid patients with substantial elbow
contracture. Thus, the indications for contracture release in the
rheumatoid population are the same as those in the general population
(that is, loss of motion impeding functional activities of daily
living). As a surgeon becomes more skilled in and knowledgeable
about a given surgical procedure, procedures that are technically
more challenging can be performed without necessarily increasing
the risk of complications. It has been well established in the medical
literature that complication rates associated with procedures ranging
from total hip arthroplasty53 to
coronary angiography54,55 vary
by surgeon and by center.
It was our purpose not only to document the compli cations of
elbow arthroscopy but also to attempt to identify factors that may
have contributed to, or prevented, their occurrence. We believe
that the single most important technical factor that prevented serious
nerve injuries was the use of retractors. We now routinely use one
or two retractors in the anterior part of the elbow when performing
synovectomy or capsulectomy. This greatly facilitates exposure.
The second factor that we believe permitted complex procedures to
be performed without serious nerve injuries was the arthroscopic
identification and, if necessary, exploration of nerves. The indication
was usually to permit capsulectomy in the regions of the nerves
(usually the ulnar and/or radial nerve) with the certainty that
the nerve was not being cut. Most of the concerns regarding complications
of elbow arthroscopy relate to the potential for nerve injury. As
such, the goal of knowing where the nerves are in order to know
where they are not is one worth striving to achieve. Our approach
was to identify the nerves arthroscopically with use of a number
of principles and techniques. First, we ensured that we were familiar
with the precise three-dimensional locations and paths of the nerves
with respect to the structures seen from an intra-articular perspective.
This was accomplished by cadaveric dissection after we had already familiarized
ourselves with these same anatomical relationships while doing open
operations. Second, the ulnar nerve often can be palpated behind
the epicondyle with a blunt probe from inside the joint and with
a finger from outside the joint. Third, once the capsule is incised,
the nerve can be gently palpated with a blunt probe. Fourth, the
nerve can be visualized. Finally, the nerve can be formally explored
as the capsular release progresses.
An alternative would be to explore the nerve through an open
approach to ensure its safety during the arthroscopic operation.
We have done this in three patients treated subsequent to this study. Along
this same line, an ulnar nerve transposition was performed first,
and then followed by the arthroscopic procedure, in eight cases
in the present series. This permitted the operation to be performed safely
while the procedure was monitored from both inside and outside the
joint. However, one must realize that even when the nerve is explored
in an open procedure, it can still slip back into the operative
field and be caught and damaged by power instruments if it is not
constantly retracted and protected.
Anterior or posterior capsulectomy should be carried out only
by surgeons who are experienced in the performance of such procedures
with use of open techniques and who also have substantial expertise
in arthroscopy. The operating surgeon must either be able to identify
the radial or ulnar nerve at the time of anterior or posteromedial
capsulectomy, respectively, or have sufficient experience with this
procedure and knowledge of the neural anatomy in relation to the
capsule and intra-articular structures so that such visualization
of the nerves is not necessary. A thorough understanding of the
three-dimensional anatomy of the elbow and surrounding nerves and
of the effects of joint distension, correct portal placement, recognition
of procedures that place the patient "at risk" for nerve injury,
and strong arthroscopic skills are necessary to prevent serious
complications of elbow arthroscopy, particularly as more complicated
procedures are performed.