Aplastic anemia is characterized by bone-marrow hypoplasia
and peripheral pancytopenia9.
It is often caused by defective stem cells and results in serious
complications such as bleeding and infection4,5.
Particularly, patients who have severe aplastic anemia according
to the guidelines of the International Aplastic Anemia Study Group
have a high mortality rate from bleeding and infection, and half
of them die within six months after the diagnosis if treated conventionally3-5,9.
Only a few investigators have reported operating on patients
with aplastic anemia12,13,20,25,27,30.
Most of the operations in these patients were confined to the abdomen.
Because far more perioperative blood loss is expected during hip
arthroplasty than during abdominal surgery, and because hip arthroplasty
is associated with the additional risk of periprosthetic infection,
special medical management is important in order to avoid complications
in patients with aseptic anemia who are managed with hip arthroplasty.
Also, the fatty marrow conversion in the proximal aspect of the femur
and the pelvis that is characteristic of the disease4 can potentially limit the durability
of the implant.
To our knowledge, no studies have been performed to specifically
evaluate the results of hip arthroplasty in patients with aplastic
anemia. The purpose of the present study was to prospectively evaluate
the perioperative management and the intermediate-term clinical
and radiographic results of hip arthroplasty in this patient population.
Between March 1990 and May 1992, twenty-six consecutive primary
hip arthroplasties were performed in nineteen patients who had been
diagnosed with aplastic anemia. There were fourteen men and five
women, and the mean age at the time of the operation was forty years
(range, twenty to sixty-six years). The indication for surgery was
osteonecrosis of the femoral head in twenty-five hips and a fracture
of the femoral neck in one. The patients were prospectively followed
with preoperative and scheduled clinical and radiographic evaluations.
The Harris hip score14 was calculated
preoperatively and at each yearly visit. The minimum duration of
clinical and radiographic follow-up was seventy-two months or until
death. No patients were lost to follow-up.
The severity of the aplastic anemia was classified according
to the criteria of Camitta et al.2-5.
The anemia was classified as severe when two of the following three
conditions were detected: a neutrophil count of less than 500 per
microliter, a platelet count of less than 20,000 per microliter, and
a corrected reticulocyte count of less than 1 percent. Ten patients
were classified as having severe aplastic anemia and nine, as having
nonsevere aplastic anemia.
Magnetic resonance imaging and histological examinations were
performed to evaluate the bone-marrow patterns of the proximal aspect
of the femur in all of the patients except for the one who had a
femoral neck fracture. The bone-marrow patterns were evaluated with
T1-weighted images, which delineate fat well (Fig. 1-A, Fig. 1-B, and Fig. 1-C). T1 relaxation
time was obtained from randomly selected sites at the junction of
the neck and the intertrochanteric region. The signal index (the
ratio of the T1 relaxation time of the neck-intertrochanteric junction
to that of the subcutaneous fat tissue) was calculated automatically
with the installed software of the magnetic resonance imaging scanner.
A signal index of 1 indicates that the T1 relaxation time of the
bone marrow of the proximal aspect of the femur and the pelvis is
identical to that of the subcutaneous fat tissue. The mean signal
index of the hips of the patients with aplastic anemia was compared
with that of a control group of hips of fifteen healthy volunteers
with use of the Student t test. Intraoperative specimens from the femoral
neck, trochanteric region, pelvis, and acetabulum were examined
histologically.
Prospective Perioperative Management
Serum platelet, hemoglobin, hematocrit, granulocyte, and corrected
reticulocyte counts were performed at three-day intervals beginning
two weeks before the operation. To prepare for very low levels of
platelets during the perioperative period, three platelet donors
were assembled before the operation to allow for platelet transfusion19,21. Five donors were needed for
one patient with platelet antibodies. A platelet concentrate of
250 to 300 milliliters (five or six units) was obtained from one
of the three donors two days before the arthroplasty and was transfused
into the patient one day preoperatively. An attempt was made to
maintain the platelet count at more than 60,000 to 70,000 per microliter.
Another 250 to 300 milliliters of platelets from the second donor
was transfused into the patient during the operation to prevent hemorrhagic
complications. In two patients with platelet antibodies, the platelet
count was kept at more than 25,000 per microliter during the perioperative
period. Postoperatively, the hematological studies were repeated
at six-hour intervals and transfusion was performed on the basis
of the amount of blood collected in the suction drain. The transfusion
was continued until the platelet and hemoglobin levels reached and
remained at or above preoperative levels.
Operative Procedure
A mean of twenty-six months (range, eight to seventy-two months)
elapsed between the diagnosis of aplastic anemia and the time of
the arthroplasty. In all cases, the posterolateral approach was
used. Three hips had a bipolar arthroplasty: one (Case 18) received
a Biomet prosthesis (Warsaw, Indiana) that was fixed with Boneloc
cement (Polymers Reconstruction A/S, Farum, Denmark), one (Case
1) received an Optifix porous-coated prosthesis (Richards Medical,
Memphis, Tennessee), and one (Case 11) received an Omnifit hydroxyapatite-coated
prosthesis (Osteonics, Allendale, New Jersey) (Table I). The remaining
twenty-three hips (sixteen patients) had a total hip arthroplasty.
Three of these hips had hybrid fixation: one (Case 13) received
a Biomet prosthesis consisting of an uncemented acetabular component
and a femoral component that was inserted with Boneloc cement, and
two (Case 15, both sides) received an Omnifit prosthesis consisting
of a hydroxyapatite-coated acetabular component and a femoral component
that was inserted with vacuum-mixed Simplex-P cement (Howmedica,
Rutherford, New Jersey). The other twenty hips had a cementless
total hip arthroplasty: five hips (Cases 2, 5, 10 [both sides],
and 17) received an Optifix porous-coated prosthesis, and fifteen received
a hydroxyapatite-coated prosthesis. Of the fifteen hips that received
a hydroxyapatite-coated prosthesis, five (Cases 4, 8 [both sides],
9, and 16) received a PSL Omnifit acetabular component (Osteonics)
and an ABG femoral component (Howmedica) and ten (Cases 3 [both
sides], 6, 7 [both sides], 12 [both sides], 14 [both sides], and 19)
received Omnifit acetabular and femoral components (Osteonics) (Table I). Press-fit
acetabular fixation without screw augmentation was performed in
four hips (Cases 2, 6, 10 [right side], and 16), and on-line fixation
with one or two screws to augment fixation was performed in nineteen.
In all cases, a twenty-eight-millimeter cobalt-chromium head was
used.
First-generation cephalosporin and gentamicin were administered
intravenously before the skin incision. The administration of antibiotics
was continued intravenously or intramuscularly for a mean of seven
days and then orally for the next fourteen days. Partial weight-bearing
was recommended for six weeks after surgery for all patients. There
were no perioperative deaths. Two patients, with three total hip
replacements (Cases 1 and 10), died at seventy-five and seventy-seven
months postoperatively, respectively. Both patients were doing well
at the time of the latest follow-up and had had no radiographic
evidence of implant loosening.
Patient Evaluation
All patients were evaluated prospectively. The Harris hip score14 was calculated preoperatively and
at each yearly follow-up visit. The hip scores of the patients with severe
and nonsevere aplastic anemia were compared with use of the Student
t test. Anteroposterior and lateral radiographs were made before
and immediately after the operation, at six weeks, at three months,
at one year, and annually thereafter. The mean duration of follow-up
was fifty-six months (range, forty-eight to seventy-one months).
Cemented hip prostheses were evaluated for loosening according
to the criteria of Harris and McGann15.
Uncemented prostheses were evaluated according to the criteria of
Callaghan et al.1.
All statistical analyses were performed with use of the Statistical
Package for the Social Sciences software (version 6.02; SPSS, Chicago,
Illinois). A p value of less than 0.05 was considered significant.
T1-weighted magnetic resonance images showed increased signal
intensity at the neck, intertrochanteric region, pelvis, and acetabulum
in all cases. The mean signal index, which represents the fat-marrow
ratio of the proximal aspect of the femur and pelvis to that of
the subcutaneous tissue, was 0.69 for the fifteen hips of the healthy
volunteers in the control group and 0.94 for the hips of the patients
with aplastic anemia (p < 0.05). These data demonstrate that
the proportion of bone marrow that was converted into fatty marrow
was greater in the hips of the patients with aplastic anemia than it
was in the normal hips. Histological examination of the intraoperative
specimens substantiated the results of magnetic resonance imaging
and showed fatty marrow conversion, increased fat-cell size, large
fat-cyst formation, mucination and degeneration of fat cells, and
compression of sinusoids.
Three days before the operation, the mean serum level of hemoglobin
was 7.8 grams per deciliter (range, 6.0 to 13.1 grams per deciliter)
(4.8 millimoles per liter [range, 3.7 to 8.1 millimoles per liter]),
the mean platelet count was 34,923 per microliter (range, 13,000
to 93,000 per microliter), and the mean neutrophil count was 1733
per microliter (range, 900 to 2800 per microliter). Two patients
(Cases 8 and 10) had platelet antibodies. After the transfusion
of platelets and whole blood one day before surgery, the mean serum
level of hemoglobin was 9.2 grams per deciliter (range, 7.2 to 12.3
grams per deciliter) (5.7 millimoles per liter [range, 4.5 to 7.6
millimoles per liter]), the mean platelet count was 73,231 per microliter (range,
23,000 to 120,000 per microliter), and the mean neutrophil count
was 1828 per microliter (range, 950 to 2950 per microliter). At
the time of surgery, the mean platelet count was more than 70,000
per microliter in fifteen patients and below this level in four.
One patient had a platelet count of less than 25,000 per microliter.
The postoperative blood loss (Hemovac drainage) averaged 656 milliliters
(range, 252 to 1274 milliliters). A mean of twenty-two units of
platelets (range, eight to fifty-four units) and 2.8 units of whole
blood (range, one to six units) were transfused in the operative and
immediate postoperative periods. Eight patients received a transfusion
of packed red blood cells (mean, three units; range, two to five
units). There were no intraoperative or postoperative hemorrhagic
complications.
The mean Harris hip score was 55 points (range, 42 to 68 points)
preoperatively and 87 points (range, 56 to 95 points) after a minimum
of six years of follow-up (Table I). At the time of the latest follow-up
evaluation, twelve hips were rated as excellent; eleven, as good;
two, as fair; and one, as poor. The poor result was for a patient
who had radiographic evidence of loosening of a bipolar prosthesis
that had been fixed with Boneloc cement. According to the criteria
of Camitta et al.2-5, ten patients
had severe aplastic anemia and nine had nonsevere aplastic anemia.
The mean Harris hip score at the time of the latest follow-up was
88 points (range, 79 to 93 points) in the eleven patients (thirteen
hips) with severe aplastic anemia and 84 points (range, 56 to 95
points) in the eight patients (thirteen hips) with nonsevere aplastic anemia.
There was no significant difference between the Harris hip scores
for these two groups either preoperatively or at the time of the
six-year follow-up. The mean Harris hip score was 74 points (56,
79, and 88 points) for the three hips treated with bipolar arthroplasty
and 89 points (range, 76 to 95 points) for those treated with total hip
arthroplasty. This difference also was not significant.
At a minimum of six years following surgery, one patient had
severe pain in the trochanteric region, groin, and thigh and another
patient had moderate pain in the groin and the anterior part of
the thigh. Two additional patients (three hips) had slight pain in
the thigh in association with weight-bearing or prolonged activity,
and one patient (one hip) had severe pain in the thigh. All of the
other patients had either no pain or slight pain. At the time of
the minimum six-year follow-up, two patients had a severe limp and
three had a moderate limp. One patient (Case 5) used a cane full-time,
and another (Case 18) always required a crutch for walking.
Some radiographic changes were seen during the follow-up period.
One patient (Case 1) was noted to have some medial protrusion of
the bipolar component at forty-eight months postoperatively. Another
patient (Case 18), who had been managed with a bipolar prosthesis
fixed with Boneloc cement because of a fracture of the femoral neck,
had definite femoral loosening (subsidence) after four years of
follow-up.
No migration of the acetabular component, subsidence of the femoral
component, radiolucency, or osteolysis occurred in the three hips
with hybrid fixation. No uncemented acetabular component migrated
during the follow-up period. Two hips with a hydroxyapatite-coated
acetabular component had progressive radiodense lines in zone III
of DeLee and Charnley8. No uncemented
femoral stem subsided during the follow-up period (Fig. 2-A, Fig. 2-B, and Fig. 2-C). One hip
had progressive retroacetabular osteolysis in zones I and II8. Three hips had areas of femoral
osteolysis, ranging from one to three centimeters in size, in zones I,
VI, and VII of Gruen et al.11.
Complications
There were three orthopaedic complications; all were intraoperative
fractures. Two of these fractures, one of which occurred in a patient
who received a hydroxyapatite-coated prosthesis (Case 6) and the
other of which occurred in a patient who received an Omnifit prosthesis
(Case 12, right side), were undisplaced femoral fractures extending
from the medial portion of the osteotomized neck to the lesser trochanter.
Neither fracture was fixed. Both patients were managed with non-weight-bearing
for six weeks and partial weight-bearing for another six weeks and
then were allowed full weight-bearing at three months. Both fractures
healed uneventfully and, at the time of the six-year follow-up,
the Harris hip scores were 90 points (Case 6) and 88 points (Case
12, right side). The third fracture, which occurred in a patient
who had hybrid fixation of an Omnifit femoral prosthesis (Case 15,
right side), was a displaced fracture extending from the tip of
the cement to the distal aspect of the femur. It occurred during
reduction of the hip after fixation with cement and was treated
with plate fixation, wiring, and six weeks of skeletal traction.
Full weight-bearing was permitted at the third postoperative month,
and the fracture healed. At the time of the six-year follow-up,
the Harris hip score was 88 points.
There were no infections or hemorrhagic complications. At the
time of this writing, one revision was planned for the treatment
of loosening of a bipolar prosthesis that had been fixed with Boneloc cement.
One patient had died of pneumonia at seventy-five months postoperatively,
and another had died of sepsis at seventy-seven months postoperatively.