Ninety-nine patients (ninety-nine hips) were managed with a Charnley total
hip arthroplasty for the treatment of a femoral neck fracture nonunion at our
institution between 1970 and 1977. The average interval between the femoral
neck fracture and the total hip arthroplasty was seventeen months (range, two
to sixty-eight months). Ninety-seven of the ninety-nine fractures had been
treated previously with internal fixation, and the other two had been treated
nonoperatively.
The study group included seventy-five women and twenty-four men and
involved forty-three right hips and fifty-six left hips. The average age of
the patients was sixty-seven years (range, thirty-two to ninety-one years) at
the time of the fracture and sixty-eight years (range, thirty-six to
ninety-two years) at the time of the arthroplasty. Sixty-one patients (62%)
were sixty-five years old or more at the time of the arthroplasty, and
thirty-eight patients (38%) were less than sixty-five years old. A cemented
Charnley cup and a cemented Charnley monoblock stem (DePuy, Warsaw, Indiana)
with a 22.225-mm head were inserted in each hip with a hand-packed cementing
technique through a transtrochanteric approach.
Patients were routinely examined at two months, one year, two years, and
five years after the operation and then at five-year intervals (or more often)
thereafter. When patients were unable to travel to a follow-up examination,
they were asked to provide clinical information by completing a standardized
questionnaire (administered by mail or by telephone) and they also were asked
to forward outside radiographs for review.
Radiographs were examined by two of us (T.M.M. and D.J.B.), neither of whom
was the operating surgeon in any case. Preoperative radiographs were reviewed
for all patients to confirm the diagnosis of femoral neck fracture nonunion.
Serial postoperative radiographs were then analyzed for acetabular component
loosening according to the criteria described by Hodgkinson et
al.12. Fixation of
the femoral component was evaluated in accordance with the categories of
loosening described by Harris et
al.13, except that
component debonding from the cement with less than a 2-mm radiolucent line at
the shoulder of the implant was not considered to represent
loosening14.
Survivorship analysis was performed with use of the Kaplan-Meier
method15. For the
end points of revision or component removal for any reason and for aseptic
loosening, patients were censored at the time of death or the time of the last
follow-up. For the end point of mechanical failure, patients were censored at
the time of death, revision, or last radiographic follow-up, whichever came
first. The log-rank test was used to assess associations between the end
points and the potential risk factors for failure, including (1) body-mass
index (<30 as compared with =30), (2) cardiopulmonary comorbidities
(severe chronic obstructive pulmonary disease, a history of myocardial
infarction, coronary artery disease, congestive heart failure, or valvular
heart disease), (3) the timing of the total hip arthroplasty (less than six
months after the fracture as compared with six months or more after the
fracture), (4) gender, and (5) age at the time of the arthroplasty (less than
sixty-five years as compared with sixty-five years or more). The level of
alpha was set at 0.05. All analyses were performed with use of the SAS
statistical software package (SAS Institute, Cary, North Carolina). The date
of death was confirmed from the Social Security Death Index.
The mortality rate for all ninety-nine patients was 0% at one year, 4% at
two years, 15% at five years, 35% at ten years, and 77% at twenty years.
Fourteen patients were still living at the time of the last follow-up.
Eleven patients were lost to follow-up less than two years postoperatively;
none of these patients had had a revision or were known to have had a failure
of the implant at the time of the last follow-up. Four other patients died
less than two years postoperatively. The remaining eighty-four patients were
followed clinically until death, revision, or implant removal or for a minimum
of two years (mean, 12.2 years; range, two to 28.9 years) postoperatively.
Twelve hips were treated with revision or component removal. Ten hips were
revised because of aseptic loosening, one hip was revised because of
dislocation, and one hip had component removal because of infection. The
femoral component alone was revised in six hips, the acetabular component
alone was revised in one, and both components were removed or revised in five.
The mean interval between the index arthroplasty and revision or implant
removal was 10.2 years (range, 0.3 to 18.7 years).
Thus, seventy-two unrevised hips were followed clinically for at least two
years. At the time of the last follow-up, forty-eight patients (67%) had no
hip pain, twenty-one (29%) had mild hip pain, and three (4%) had moderate or
severe hip pain. At the time of the last follow-up, twenty-four patients (33%)
walked with no arm support, twenty-seven (38%) walked with one arm support,
thirteen (18%) walked with two arm supports, and eight (11%) were unable to
walk. None of the patients who were unable to walk had moderate or severe hip
pain.
Radiographs that had been made at least two years postoperatively were
available for review for forty-seven unrevised hips; the average duration of
radiographic follow-up for these hips was 12.6 years (range, 2.4 to 25.3
years). Sixteen (34%) of these forty-seven hips had radiographic signs of
loosening: eight had loosening of the cup, six had loosening of the stem, and
two had loosening of both components.
Dislocation occurred in nine patients (9%) and was the second most common
complication after loosening. Four patients had a single dislocation, and five
patients had two dislocations or more. Only one hip was revised because of
this complication, as discussed above.
The survival rate free of revision or implant removal for any reason was
97% (95% confidence interval, 93% to 100%) at five years, 93% (95% confidence
interval, 87% to 99%) at ten years, and 76% (95% confidence interval, 63% to
91%) at twenty years (Fig. 1).
The survival rate free of component revision for aseptic loosening was 95%
(95% confidence interval, 90% to 100%) at ten years and 78% (95% confidence
interval, 64% to 93%) at twenty years. The survival rate free of mechanical
failure (as indicated by either radiographic signs of loosening or revision
for aseptic loosening) was 97% (95% confidence interval, 93% to 100%) at five
years, 84% (95% confidence interval, 74% to 95%) at ten years, 62% (95%
confidence interval, 48% to 79%) at fifteen years, and 39% (95% confidence
interval, 23% to 63%) at twenty years.
The survival rate free of revision for mechanical failure was better in
patients who had been sixty-five years old or more at the time of the
operation than it was in those who had been less than sixty-five years old (p
= 0.045). The implant survival rate for patients who had been less than
sixty-five years old at the time of the arthroplasty was 94% (95% confidence
interval, 87% to 100%) at ten years and 65% (95% confidence interval, 46% to
90%) at twenty years, whereas the implant survival rate for patients who had
been sixty-five years old or more at the time of the arthroplasty was 95% (95%
confidence interval, 89% to 100%) at ten years and 95% (95% confidence
interval, 83% to 100%) at twenty years
(Fig. 2). For the same end
points, a body-mass index of =30 (p < 0.01) and male gender (p = 0.02)
also correlated with poorer survival rates. With the numbers available,
neither the time-interval between the initial femoral neck fracture and the
total hip arthroplasty nor the presence of serious cardiac or pulmonary
comorbidities was correlated with the implant survival rate with revision for
any reason, revision for aseptic loosening, or mechanical failure as the end
point (p > 0.05).
When nonunion of a femoral neck fracture occurs in an older patient,
replacement of the femoral head by means of a hip arthroplasty has been
advocated because it obviates further problems with bone-healing and allows
for the rapid mobilization of the patient. Shorter-term results in several
smaller series have demonstrated the efficacy of this procedure in reducing
pain and improving
function4,6,10,11.
However, we are aware of no study that has included a sufficient number of
patients who were followed for a sufficient period of time to assess the
durability of this operation in this population or to stratify how
patient-related factors, such as age at the time of the operation, affect
results.
The present study also demonstrated that the procedure can provide good
early clinical results. The longer-term durability of the implant was
reasonably good in older patients (those who had been sixty-five years old or
more at the time of the operation), but it was less good in younger patients,
in men, and in patients with a high body-mass index. While we acknowledge the
limitations of comparing the results of the present study with those of
previous studies involving different groups of patients, we found that the
overall implant survival rate in the current report was somewhat poorer than
those in studies of Charnley total hip arthroplasty in the general
population16,17,
despite the older mean age of our patients (sixty-eight years) and the
predominance of women in the current study, both of which are factors that
would have been expected to improve
survival18.
Kavanagh et al.16
reported a twenty-year survival rate of 84% in their study of Charnley total
hip arthroplasties in the general population, compared with the rate of 76% at
the same time-interval in the present study. One may speculate that Charnley
total hip arthroplasty could be less durable in patients with a nonunion at
the site of a femoral neck fracture because such patients have poorer
acetabular and femoral bone quality (by virtue of the fracture and subsequent
nonunion) and because they frequently have holes in the femur at the sites of
old fixation devices. Each of these factors could reduce the durability of
cemented implants, especially when so-called first-generation hand-packed
cementing techniques are used. Finally, a subgroup of patients with this
diagnosis comprises younger patients with unilateral hip disease, who are at
risk for mechanical implant failure.
A notable complication in the present study was dislocation of the hip,
which occurred in 9% of the patients. A high dislocation rate has been
demonstrated in other series in which total hip arthroplasty was performed for
the treatment of nonunion at the site of a femoral neck
fracture8,19.
The present study demonstrates the efficacy of total hip arthroplasty with
cement fixation for the treatment of nonunited femoral neck fractures, but it
also demonstrates problems with durability in association with the use of
first-generation cementing techniques, particularly (but not exclusively) in
younger patients. On the basis of these results, we believe that young
patients with this problem may be considered as candidates for joint-sparing
procedures or for arthroplasty with more modern implants and surgical
techniques. When arthroplasty is chosen for the treatment of this complication
in older patients, efforts to optimize implant fixation with use of more
advanced arthroplasty techniques with or without cement appear to be
justified. Because dislocation was the most common complication aside from
implant loosening, efforts to minimize the risk of dislocation by optimizing
the choice of operative approach, implant, and operative technique also are
advocated.