Between 1975 and 2000, eighty-three femoral neck fractures in eighty-two
consecutive patients with a mean age of thirty-six years (range, fifteen to
fifty years) were treated with internal fixation at our level-1 trauma center.
All fractures were classified as type 31B according to the system of the
Orthopaedic Trauma
Association14. All
patients were skeletally mature, and those who had pathologic fractures due to
neoplasm were excluded. Fifty-three patients were male, and twenty-nine were
female. Fifty-nine fractures were displaced, and twenty-four were
nondisplaced. Two patients died, and eight (9.8%) were lost to follow-up.
Seventy-three fractures in seventy-two patients were followed until union,
until conversion to hip arthroplasty, or for a minimum of two years. The mean
duration of clinical follow-up for the group as a whole was 6.6 years (range,
three months to twenty-three years). Ten patients demonstrated fracture union
but did not subsequently return for follow-up. These ten patients had been
followed for a mean of ten months (range, three months to 1.3 years). Five
patients required conversion to total hip arthroplasty after less than two
years of follow-up. Therefore, fifty-seven patients with fifty-eight fractures
were followed for a minimum of two years (mean, 8.1 years; range, two to
twenty-three years).
Follow-up radiographs were available for fifty-four (74%) of the
seventy-three fractures. The radiographs for the other nineteen fractures had
been discarded as part of institutional policy or lost. In those cases,
however, the notes of the radiologist and the treating orthopaedic surgeon
clearly documented the outcome with respect to fracture-healing and femoral
head osteonecrosis; thus, these fractures were included in the study. The mean
duration of radiographic follow-up for the group as a whole was 5.4 years
(range, three months to twenty-three years). Fifty-one of the fifty-seven
patients with a minimum of two years of clinical follow-up had a minimum of
two years (mean, 7.4 years; range, two to twenty-three years) of radiographic
follow-up.
Clinical and radiographic data were retrospectively reviewed, and the
results and complications were analyzed. Institutional review board approval
was obtained for a retrospective review. In polytraumatized patients,
life-threatening injuries were treated first and then the femoral neck
fractures were treated in an urgent fashion. Fifty-three fractures were
treated within twenty-four hours after diagnosis, and twenty were treated
after more than twenty-four hours because of late presentation (fifteen) or
the need to treat other life-threatening injuries (five). The mechanism of
injury was a fall for thirty-five patients (thirty-five fractures), a
motor-vehicle accident for twenty-seven patients (twenty-eight fractures), and
a stress or insufficiency fracture with no documented single traumatic event
for ten patients (ten fractures).
Fifty-one fractures were displaced, and twenty-two were nondisplaced.
According to the system of the Orthopaedic Trauma Association, there were
thirty-seven transcervical (type-31B2) fractures, twenty-six subcapital
(type-31B1) fractures, and ten basicervical (type-31B2.1) fractures. The
choice of fixation device, the operative approach, and the need for
capsulotomy were determined by the treating surgeon. Fifty-two fractures were
treated with cannulated screws; seventeen, with a sliding hip screw; two, with
a reconstruction nail; one, with an angled blade-plate; and one, with a
McLaughlin plate.
Fourteen of the fifty-one displaced fractures were treated with open
reduction (with direct visualization of the fracture fragments) and internal
fixation, and thirty-seven were treated with closed reduction and internal
fixation. Four of the thirty-seven displaced fractures that were treated with
closed reduction also were treated with capsulotomy. Three of the twenty-two
nondisplaced fractures were treated with capsulotomy, and one was treated with
aspiration of the hip.
Thirty-one (43%) of the seventy-two patients had seventy associated
injuries, including an ipsilateral femoral shaft fracture (fifteen patients;
21%), a contralateral femoral fracture (two), a tibial or ankle fracture
(ten), an upper extremity fracture (seven), a closed head injury (seven), a
patellar fracture (six), a pelvic fracture (six), a foot fracture (five), an
acetabular fracture (four), a rib fracture (two), and a skull fracture, facial
fracture, cervical spine fracture, liver laceration, renal artery occlusion,
and duodenal transection (one each).
Because there is no generally accepted method of grading the quality of
reduction of a femoral neck fracture, fracture reduction was graded on the
basis of the degree of residual angulation and the amount of displacement as
excellent (<2 mm of displacement and <5° of angulation in any
plane), good (2 to 5 mm of displacement and/or 5° to 10° of
angulation), fair (>5 to 10 mm of displacement and/or >10° to
20° of angulation), or poor (>10 mm of displacement and/or >20°
of
angulation)5,15-17.
Nonunion was defined as failure of fixation with implant breakage, loss of
reduction, or persistence of a visible fracture line at a minimum of six
months after the index procedure. Osteonecrosis was classified
radiographically with use of the method of Ficat and
Arlet18. Functional
assessment at the time of follow-up was performed by evaluating pain, walking
status, and the need for gait aids as documented by the treating surgeon.
Survival of the native femoral head was calculated with use of the
Kaplan-Meier method, with conversion to total hip arthroplasty as the end
point.
For the group as a whole, fifty-three (73%) of the seventy-three fractures
healed after one operation and demonstrated no evidence of osteonecrosis at
the time of the last follow-up. Seventeen fractures (23%) were associated with
the development of osteonecrosis, which was classified as stage I in one hip,
stage II in four, stage III in five, and stage IV in seven. Six fractures (8%)
were associated with the development of nonunion. Three fractures, therefore,
were associated with the development of both osteonecrosis and nonunion
(Table I).
Five (9.8%) of the fifty-one displaced fractures were associated with the
development of nonunion, and fourteen (27%) were associated with the
development of osteonecrosis. Three displaced fractures were associated with
the development of both osteonecrosis and nonunion. Three (14%) of the
twenty-two nondisplaced fractures were associated with the development of
osteonecrosis, and one (4.5%) was associated with the development of nonunion.
There was a strong trend for displaced fractures to demonstrate higher rates
of osteonecrosis and nonunion; however, with the numbers available, the
difference was not significant (p = 0.17).
At the time of the most recent follow-up, thirteen (18%) of the
seventy-three fractures had been treated with a total hip arthroplasty because
of the development of osteonecrosis (eleven), nonunion (one), or both (one).
The mean time to conversion to total hip arthroplasty from the time of the
injury was 7.3 years (range, three months to fifteen years). The rate of
survival of the native femoral head free of conversion to total hip
arthroplasty was 88.4% (95% confidence interval, 80.3% to 97%) at five years
and 85% (95% confidence interval, 74.1% to 95.9%) at ten years
(Fig. 1). Five other patients
had additional operations for the treatment of nonunion or osteonecrosis.
Specifically, four additional procedures (two procedures involving a quadratus
femoris muscle-pedicle bone graft, one valgus-producing intertrochanteric
osteotomy, and one repeat open reduction and internal fixation with autologous
bone-grafting) were performed for the treatment of nonunion, and one
additional procedure (a repositional proximal femoral osteotomy) was performed
for the treatment of osteonecrosis. All four secondary attempts to achieve
fracture union were successful; however, one patient in whom successful union
was achieved later required total hip arthroplasty for the treatment of
osteonecrosis. Therefore, seventy-one (97%) of seventy-three fractures
ultimately demonstrated osseous union.
All nondisplaced fractures, by definition, had an excellent reduction.
Therefore, in the group as a whole, sixty-eight (93%) of the seventy-three
fractures had a good to excellent reduction. Of these, fourteen (21%) were
associated with the development of osteonecrosis and three (4%) were
associated with the development of nonunion. Of the forty-six displaced
fractures with a good to excellent reduction, eleven (24%) were associated
with the development of osteonecrosis and two (4%) were associated with the
development of nonunion. Five displaced fractures (10%) had a fair or poor
reduction. Two fractures were associated with the development of both
osteonecrosis and nonunion, one was associated with the development of
osteonecrosis, and one was associated with the development of nonunion. Thus,
only one of five fractures that had a fair or poor reduction healed without
complication.
Fourteen displaced fractures were treated with open reduction, with direct
visualization of the fracture fragments. These fractures, by definition, were
treated with capsulotomy. Four additional fractures that were treated with
successful closed reduction and internal fixation also were treated with
capsulotomy. Three nondisplaced fractures were treated with capsulotomy, and
one was treated with aspiration of the hip. Therefore, for the group as a
whole, twenty-two (30%) of the seventy-three fractures were treated with some
form of capsular decompression and fifty-one of the seventy-three fractures
were not. At the time of the most recent follow-up, four (18%) of the
twenty-two fractures that had been treated with capsular decompression and
thirteen (25%) of the fifty-one fractures that had been treated without
decompression were associated with the development of osteonecrosis. With the
numbers available, this difference was not significant (p = 0.50).
Additionally, we could not demonstrate a significant difference when analyzing
the impact of capsular decompression on the rate of osteonecrosis for
displaced fractures (p = 0.33) and nondisplaced fractures (p = 0.47).
Of the fifty-three fractures that were treated within twenty-four hours
after diagnosis, thirteen (25%) were associated with the development of
osteonecrosis and four (7%) were associated with the development of nonunion;
two of these fractures were associated with the development of both. Of the
twenty fractures that were treated more than twenty-four hours after
diagnosis, four (20%) were associated with the development of osteonecrosis
and two (10%) were associated with the development of nonunion; one of these
fractures was associated with the development of both. With the numbers
available, the difference was not significant.
Two patients had an intraoperative complication. In one patient an
intra-articular screw necessitated a return to the operating room for screw
repositioning, and in one patient a broken drill-bit was left in the proximal
part of the femur. Twelve (17%) of the seventy-two patients had a
postoperative complication; the complications included pneumonia (four
patients), wound infection requiring incision and drainage and delayed closure
(two), deep venous thrombosis (one), sepsis (one), urinary tract infection
(one), ileus (one), thrombocytopenia (one), and atrial fibrillation (one).
At the time of the last follow-up, the fifty-nine patients (sixty
fractures) with a preserved femoral head generally had excellent clinical
function: fifty-eight fractures (97%) caused no or mild pain (that is, pain
that was associated with no activity restrictions and no use or occasional use
of nonnarcotic analgesics), and two caused moderate pain (that is, pain that
was associated with activity limitations and regular use of nonnarcotic
analgesics). The great majority of patients did not require gait aids and
reported no activity restrictions: only seven used a cane occasionally, and
only one required two-arm support (secondary to traumatic brain injury and
hemiparesis). One patient was unable to walk because of other injuries. In the
subset of five patients with osteonecrosis who had not undergone conversion to
total hip arthroplasty, three reported no pain and two reported mild pain. One
patient occasionally used a cane; the rest were unrestricted and used no gait
aids.
The present study demonstrated that the use of contemporary internal
fixation methods for the primary treatment of femoral neck fractures in young
patients was associated with a high rate of fracture union (92%; sixty-seven
of seventy-three). When secondary procedures were considered, rate of healing
was even higher (97%; seventy-one of seventy-three). However, even with the
use of contemporary treatment methods, seventeen (23%) of seventy-three
fractures were associated with the development of osteonecrosis.
Osteonecrosis, not nonunion, was the most common problem leading to conversion
to hip arthroplasty. The overall rate of femoral head retention was 82% at a
mean of 6.6 years. Although this finding is encouraging, 18% of our patients
required conversion to hip arthroplasty at a young age, which demonstrates the
need for continued efforts to optimize the treatment of this injury. The high
rate of fracture-healing in these patients probably was due to the healing
potential and good bone quality of the femoral head and neck of most young
patients19,20.
In contrast, older patients with femoral neck fractures have poorer bone
quality, which probably leads to higher rates of nonunion; a rate of >30%
was reported in a recent
meta-analysis12.
The 27% rate of osteonecrosis among patients with displaced fractures is
similar to that reported in previous, smaller studies of younger patients with
displaced fractures, even studies in which excellent reduction was achieved in
a high percentage of patients and capsulotomy was used
routinely5-7,9-11,13,16.
Fortunately, a substantial number of the patients who had osteonecrosis in the
present series (five [29%] of seventeen) did not have marked symptoms and had
not required additional surgery at the time of the latest follow-up. In a
recent series by Jain et
al.10, the
occurrence of osteonecrosis did not significantly affect functional outcome;
however, the mean duration of follow-up for the entire group was only 2.5
years.
The findings of the current study support the premise that the quality of
reduction has an impact on the outcome of treatment. The outcome for patients
with a fair or poor reduction was poor, although it is important to recognize
that problems that make it difficult to obtain a good reduction, such as
comminution and marked displacement, may reflect more severe injury patterns.
It is also important to note that although most patients had a good to
excellent reduction, approximately one in five still had development of
osteonecrosis. It has been speculated that the fate of the femoral head is
partly determined at the time of injury, and in many patients this is probably
the case11. On the
basis of current data, we recommend continued efforts to obtain as accurate a
reduction as possible. With the numbers available, however, we were not able
to identify any correlation between the fracture classification or the method
of internal fixation and the rate of osteonecrosis or nonunion.
The role of capsulotomy remains
controversial21-25.
The decision to perform a capsulotomy or hip aspiration was made by the
treating surgeon. Because of the retrospective nature of this review and some
selection bias (resulting from the fact that patients with displaced fractures
were more likely to have a capsulotomy to facilitate reduction), we were
unable to clearly determine the impact of capsulotomy or hip aspiration on
outcome. Although the numbers in the current study were too small for us to
draw definitive conclusions, we did not detect a significant difference in the
rate of osteonecrosis between fractures that had been treated with a
decompressive maneuver and those that had not. The rate of osteonecrosis in
the current series is very similar to the rates in previous studies in which
capsulotomy was routinely
performed5,6,13.
Additionally, in a recent series of thirty-eight patients under the age of
sixty years, the reported rate of osteonecrosis was 16% overall, with only one
patient undergoing capsular
decompression10.
Until a large, prospective, randomized series is available, however, we will
continue to perform capsulotomy because it adds minimal additional risk, it is
simple to perform, and it may theoretically help a small subset of patients by
decompressing the blood vessels supplying the femoral head.
The shortcomings of the present study include the use of retrospective
methodology, the involvement of multiple surgeons using different fixation
devices, the inconsistent use of capsulotomy, and the fact that these
relatively rare fractures were treated over a period of many years during a
time when surgical techniques and internal fixation devices were evolving. It
is possible that complication rates may decrease further with the development
of improved surgical techniques and internal fixation devices. In addition, a
validated system was not used to describe patient outcomes.
The strengths of the present study include the large number of consecutive
patients who had been treated at a single institution and the high rates of
clinical and radiographic follow-up, which allowed us to accurately determine
the rates of osteonecrosis, nonunion, and femoral head survival in a
population of trauma patients after a relatively long duration of follow-up.
The limitations of the current study do not undermine its most important
findings, which were that (1) the ten-year survivorship of the native femoral
head free of conversion to total hip arthroplasty after femoral neck fracture
was 85% and (2) the functional outcome of the vast majority of patients was
quite good. Despite contemporary techniques of stable internal fixation and
the high percentage of good to excellent reductions, osteonecrosis developed
in >20% of patients and nonunion developed in 8%. Even though a substantial
subset of patients who had osteonecrosis were asymptomatic or minimally
symptomatic, osteonecrosis was the main reason for conversion to total hip
arthroplasty. In most patients with nonunited fractures, union was achieved
after a secondary procedure. The results of treatment were influenced by
fracture displacement and the quality of reduction. The role of capsulotomy
remains controversial.