Intramedullary nailing is the treatment of choice for displaced tibial
shaft fractures in adults
1-6
. Chronic anterior knee pain at the insertion site of the nail is the
most frequently reported complication of closed nailing
1,3,4,7-12
. Up to 56% of patients note some degree of chronic knee pain
7
.
The etiology of anterior knee pain after nailing is still unknown7-10,12.
Some investigators have proposed that a transpatellar tendon approach
for nail insertion is associated with a higher prevalence of anterior
knee pain than is a medial paratendinous approach
10,12
.
The purpose of this randomized, prospective study was to determine
if there is a difference in the prevalences of chronic anterior
knee pain following these two surgical approaches.
Fifty consecutive patients in whom an isolated, displaced fracture
of the tibial shaft was treated with an intramedullary locking nail
(Grosse-Kempf nail; Howmedica, Rutherford, New Jersey) at Tampere
University Hospital, Finland, between July 1996 and January 1998,
were enrolled in this study. There were twenty-three men and twenty-seven
women with a mean age (and standard deviation) of 42 ±;
5 years. Inclusion criteria included a patient age of fifteen years or
more, the absence of any major comorbid illness, and the absence
of any fracture lines extending up to the knee joint. All patients
were informed of the study procedure, purposes, and known risks,
and all gave written informed consent. The study was conducted in
conformity with the principles of the Declaration of Helsinki and
was approved by our institutional ethics committee. After they gave
informed consent, the patients were randomized, with use of sealed
envelopes, into one of two groups: closed nailing with use of a
patellar tendon-splitting (transtendinous) approach (twenty-five
patients) or closed nailing with use of a paratendinous approach (twenty-five
patients). One patient in each group was excluded from the study
because the nail had to be removed in the early phase of the treatment
as a result of a deep infection. Another patient in each group died
from causes not related to the tibial shaft fracture. In addition,
we were unable to contact four patients (two in each group) for
the follow-up examination. Thus, twenty-one patients in each group
were available for long-term follow-up.
The intramedullary nailing was performed within twenty-four hours
after the injury in forty-two patients and between one and twelve
days after the injury in six (two patients treated with the transtendinous
approach and four treated with the paratendinous approach). Two
patients (one in each group) were managed initially with a cast
and then were treated with corrective osteotomy and intramedullary
nailing (one, at six months and the other, at eighteen months after
the injury).
The incision for the transtendinous approach was made longitudinally
through the midline of the tendon for a distance of 3 to 5 cm. Following
insertion of the nail, the tendon incision was closed with interrupted
suture. For the paratendinous approach, a medial longitudinal incision
was made with care taken not to violate the patellar tendon or its
sheath. The entry portal in the bone was made immediately behind
the patellar tendon in all patients. Proximal and distal locking
screws were always used, and all nails were countersunk below the cortical
bone of the proximal part of the tibia. Nails were routinely removed,
through the same entry incision and approach that were used during
nail insertion, approximately one and a half years after fixation.
The forty-two patients returned for final evaluation at an average
of 1.7 ±; 0.3 years after nail extraction and 3.2 ±;
0.4 years after nail insertion. The patients graded anterior knee
pain during rest, walking, running, squatting, kneeling, stair-climbing,
and stair descent and after long-term sitting on a 100-mm visual-analog
scale, with 0 denoting no pain and 100 denoting the worst pain that
the patient could imagine
13
. Three patients treated with the transtendinous approach and four
patients treated with the paratendinous approach had not started
running at the time of the follow-up examination and thus could
not grade pain during running. The patients also assessed impairment
caused by the anterior knee pain with use of a 100-mm scale, on
which 0 meant no impairment, <33 meant mild impairment, 33 to
66 meant moderate impairment, and >66 meant severe impairment. In
addition, all patients completed the standardized scoring scales
described by Lysholm and Gillquist
14
and by Tegner et al.
15
as well as the Iowa knee score
16
.
One physician performed a blinded reexamination of all patients.
The functional evaluation was performed with use of a modification
of the method developed by Kannus et al.
17
. The evaluator used a 0 to 3-point scale to rate the patients' ability
to perform one-leg jumping and duck-walking (3 points signified
the ability to perform without problems and no pain and 0 points
signified the inability to perform and intense pain), their ability
to perform a twenty-five-repetition full-squat test (0 signified
the inability to perform any squat without pain; 1 point, the ability
to perform one to ten squats without pain; 2 points, eleven to twenty
squats; and 3 points, more than twenty squats), and their ability
to kneel (0 points meant that it was impossible to kneel; 1 point
meant that it was possible to kneel for less than ten seconds without
pain; 2 points, for less than twenty seconds; and 3 points, without time
limitation). The scores acquired from these tests were summed and
averaged for the final between-groups comparison.
The impairment scale and the functional testing scale were developed
specifically for the purpose of this study. For validation, we calculated
the Spearman correlation coefficients between our impairment scale
and the Lysholm score
14
(-0.305; p = 0.050), between our impairment scale and the standardized
Iowa knee score
16
(-0.610; p < 0.001), between our functional testing scale and the
Lysholm
14
score (0.688; p < 0.001), and between our functional testing scale
and the Iowa knee score
16
(0.734; p < 0.001).
To describe the data, the mean (and standard deviation) and the
95% confidence interval of the differences between the groups are
reported for continuous variables. Percentages as well as the risk
ratio and its 95% confidence interval are reported for categorized
variables.
In our statistical analyses, differences between the two groups were
tested with use of the Mann-Whitney U test for continuous variables
and the Fisher exact test for categorized variables. The preliminary
power calculations suggested that, with use of <5% probability
of a type-I error (p < 0.05) and a power of 80% (type-II error,
0.20), a sample size of thirty patients (fifteen patients in each
group) was necessary to detect a 50% difference in the overall success
rate between the groups. With use of an alpha level of <5% (p < 0.05)
and with a 50% difference in the overall success rate between groups considered
clinically relevant, the final number of twenty-one patients in
each group resulted in a minimum power of 93% for the study.
The statistical analysis was carried out with use of the SPSS for
Windows program (version 10.0; SPSS, Chicago, Illinois). Throughout
the study, a p value of <0.05 was considered significant.
The study groups were comparable with respect to sex, age, body-mass
index, nail protrusion, nail-plateau distance, and the durations
of follow-up after nailing and after nail removal (
Table I
). Two patients (one treated with the transtendinous approach and
one treated with the paratendinous approach) refused to have the
nail removed because of the absence of any symptoms related to it.
No patient had complications, such as infection, patellar tendon
rupture, or broken hardware, that might have contributed to any
knee pain.
Eighteen (86%) of the twenty-one patients treated with the transtendinous
approach and seventeen (81%) of the twenty-one patients treated
with the paratendinous approach reported anterior knee pain before
nail removal (p = 1.000, relative risk = 1.06, 95% confidence interval
= 0.81 to 1.39). Twelve of the eighteen patients treated with the
transtendinous approach and eleven of the seventeen treated with
the paratendinous approach had complete or marked resolution of
the pain after nail removal. The remaining six patients in each
group reported no change in the pain or worsening after nail removal.
At the time of final follow-up, fourteen (67%) of the twenty-one
patients treated with the transtendinous approach and fifteen (71%)
of the twenty-one treated with the paratendinous approach had anterior
knee pain during one or more of the activities that they assessed
with the visual analog scale (p = 1.000, relative risk = 1.07, 95%
confidence interval = 0.71 to 1.61). Thirteen of the fourteen patients
treated with the transtendinous approach reported that they experienced
pain at the patellar tendon, whereas the remaining patient experienced
pain at the medial border of the patellar tendon. None of these
patients had tenderness on the lateral side of the patellar tendon.
Seven of the fifteen patients treated with the paratendinous approach
reported pain at the patellar tendon, seven had pain at the medial
border of the tendon, and one had pain at the lateral border of
the tendon.
With the numbers available, no significant difference was found
between the study groups with respect to the prevalence and intensity
of anterior knee pain during rest, walking, running, squatting,
kneeling, or stair-climbing or descent or after long-term sitting
(
Table II
). Thirteen of the fourteen patients treated with the transtendinous
approach and ten of the fifteen patients treated with the paratendinous
approach (p = 0.536) who experienced anterior knee pain reported
that this pain caused mild-to-severe impairment during daily activities.
Five of the thirteen patients treated with the transtendinous approach
and two of the ten patients treated with the paratendinous approach
who had impairment reported that it was moderate, and one patient in
each group reported that it was severe (p = 0.669).
Compared with the score before the injury, the mean Tegner score
at the time of follow-up decreased 0.52 ±; 1.21 points
for the patients treated with the transtendinous approach and 0.19 ±;
0.81 point for those treated with the paratendinous approach (p
= 0.301, between-groups difference = -0.33 point, 95% confidence
interval = -0.98 to 0.31). The mean follow-up Lysholm activity score
was 90.4 ±; 13.9 points for the patients treated with
the transtendinous approach and 92.1 ±; 13.7 points for
those treated with the paratendinous approach (p = 0.698, between-groups
difference = -1.67 points, 95% confidence interval = -10.28 to 6.95).
The Iowa knee scores were 95.4 ±; 6.5 and 96.1 ±;
8.7 points, respectively (p = 0.765, between-groups difference =
-0.71 point, 95% confidence interval = -5.50 to 4.08).
Anterior knee pain was reported by twenty-nine (69%) of our forty-two
patients at an average of one and a half years after nail removal.
This prevalence is higher than that reported in previous retrospective
studies
3,4,7,10,12
. Seventy-nine percent of our patients who had anterior knee pain
also had some degree of impairment attributable to this pain, although
only two patients reported severe impairment. This high prevalence
should be of clinical concern.
With the numbers available, we could not find any association between
the entry incision and anterior knee pain. In their retrospective
studies, Keating et al.
10
and Orfaly et al.
12
found a clear association between a transtendinous surgical approach
and chronic anterior knee pain, and they recommended the routine
use of a medial paratendinous approach. In contrast, Court-Brown
et al. did not find any association between the surgical approach
and anterior knee pain7.
There are many factors other than the surgical approach that may
cause anterior knee pain after intramedullary nailing of a tibial
shaft fracture. Some authors have identified younger patients as
being at greater risk for chronic anterior knee pain
7,10
. This observation may be attributable to the more sedentary lifestyle
of elderly patients. Nail prominence above the cortex of the proximal
part of the tibia may be a contributing factor. However, although
excessive nail prominence clearly irritates the overlying patellar
tendon, Keating et al. found no association between nail protrusion
and anterior knee pain
10
. We also noted no relationship between a few millimeters of nail
prominence and anterior knee pain in our small series of patients.
Hernigou and Cohen suggested that, in some patients, the cause of
the anterior knee pain may be a torn meniscus or an unrecognized
articular injury
9
. Devitt et al. reported that the contact pressure in the patellofemoral
articulation increased after nailing regardless of the approach
that had been used
8
. J�rvel� et al. noted that the most important factor related to the
occurrence of anterior knee pain after reconstruction of the anterior
cruciate ligament with bone-patellar tendon-bone graft was an extension
torque deficit
18
. We found an extension torque deficit of the thigh muscles after
tibial nailing, but it was not associated with the type of entry
point that had been used. We speculate that additional causes for
anterior knee pain may be damage to the infrapatellar nerve or surgically
induced scar formation.
Nail removal partially lessens anterior knee pain
7,10
. However, the results of our study showed that the majority of patients
continued to have anterior knee pain even after removal of the nail.
Since previous investigators reported the intensity of the anterior
knee pain to be only slight or moderate
7
, the effect size of 50% was considered to be clinically relevant
in this study. Because of the relatively small sample size, the
power of our study would decrease if a smaller effect size was regarded
as relevant. For example, if a 40% or 30% effect size were selected,
the power of this study would be reduced to 76% or 50%, respectively.
We concluded that it is not possible to reduce anterior knee pain
by using a paratendinous approach rather than a transtendinous incision
for closed nailing of tibial shaft fractures. Although chronic anterior
knee pain occurred in the majority of our patients, it was rarely
severe. Additional studies to assess the role of other factors in
chronic anterior knee pain are warranted.