Introduction
The purpose of this investigation was to determine the incidence of false positive and false negative interpretations of pseudarthrosis by evaluating anteroposterior roentgenograms, lateral bending roentgenograms, and lumbar computerized tomography (CT) with multi planar reformations (MPR) in previously fused patients.
Materials and Methods
From January 1987 to April 1989 anterior interbody lumbar fusion (AILF)
was performed on a series of 78 intervertebral disks in 39 patients for
failed back surgery syndrome (FBSS), ranging from two to six surgeries.
All patients had undergone previous posterolateral fusion (PLF), which may
or may not have had diagnostically identifiable pseudarthrosis. These patients
were selected to evaluate clinical and surgical motion of each intervertebral
disk space incorporated in the PLF. The only levels within the fusion reviewed
were those ultimately fused by an AILF. Some levels above and below the
AILF may have had a previous PLF. but were not studied because motion could
not be determined without first performing an anterior lumbar diskectomy.
Many of these patients underwent an anterior and posterior surgery for multiple
reasons. Some patients had augmentation of the PLF and decompression of
foramens or central canals before the AILF was initiated. There were 16
males and 23 females, and the patients ranged in age from 20 years to 69
years (mean age, 41 years). All but three patients had the PLF more than
1 year before undergoing the AILF. The remaining three patients had the
PLF more than 9 months before the AILF.
In preoperative computer tomography (CT) scans with multiplanar reformations
(MPR), lateral bending and plain anteroposterior (AP) roentgenograms were
reviewed in blind fashion by WS, a coauthor and Radiologist. A total of
28 patients had CT with MPR, 36 patients had sidelying lateral bending films,
30 patients had AP films, and 19 patients had all three preoperative studies
performed to review for pseudarthrosis. Clinical radiographic review was
performed at each motion segment of the fusion mass, so each segment could
be categorized into an apparent fused segment or an apparent pseudarthritic
segment. The fusion mass by CT scanning was evaluated by reviewing all films,
especially curved coronal and sagittal reformations. If a pseudarthrosis
was suspected on both sides of the PLF in any view, the level was recorded
as the pseudarthrosis. If the pseudarthrosis was seen on only one side,
then the fusion mass was considered solid. Similarly, the AP roentgenograms
needed bilateral evidence of a pseudarthrosis to be considered positive
for a pseudarthrosis. If a solid mass was seen unilaterally, the fusion
mass was considered fused. Likewise, the bending films were also reviewed
at each level and if any motion was seen, the segment was considered not
fused. To review the bending films, the flexion and extension films were
superimposed to identify any motion.
All patients underwent an AILF because of persistent postoperative pain.
Some also required a posterior procedure before the AILF. An AILF was performed
if discography reproduced the patient's moderate to severe concordant back
and/or leg pain, even if the patients appeared to have a solid fusion from
clinical studies. The technique was routinely performed by us, and a vascular
surgeon performed the retroperitoneal approach. A rectangular window was
placed in the intervertebral disk space to prepare a bed for perforated
demineralized cortical femoral ring (PDCR). Anterior diskectomy and endplate
decortication were performed, followed within the appropriate sized ring
to complete the procedure. To determine motion at each intervertebral disk
space, a kidney rest, elevated to a height of 12 inches, was used to increase
lordosis in the lumbar spine. The kidney rest was then lowered to observe
motion. An intraoperative finding of motion exceeding 2 mm, measured by
calipers, was considered to be pseudarthrosis. This amount of motion corresponds to at least 3° of intervertebral angulatory change.
Results
The distribution of fusion levels evaluated by motion at the time of
the AILF are recorded in Table 1. There were 8 patients evaluated
for an L3S1 PLF, 21 patients evaluated for an L4S1 PLF, 6 patients evaluated
for an L5S1 PLF, 1 patient for an L2L4 PLF, 1 patient for an L3L5 PLF, and
2 patients for an L4L5 PLF.
We reviewed 78 levels in 39 patients to determine if any correlation could be found between the intraoperative findings of motion at each intervertebral level and the diagnostic studies (i.e., CT scan with MPR, lateral bending films, and AP films). Referring to Table 2, 54 levels were reviewed by CT scanning. There was a 42% inaccuracy rate in determining the preoperative status of the fusion mass. We found 17 levels (31%) had false negative pseudarthrosis and 6 levels (11%) had false positive pseudarthrosis. Of the 54 levels, 37 levels (68% ) were determined to have intraoperative motion greater than 2 mm, and 17 levels had no intraoperative motion.
The bending films showed a 36% inaccuracy rate of diagnosing a pseudarthrosis. There were 20 (30%) false negative pseudarthroses and 4 (6%) false positive pseudarthroses. There were 67 levels reviewed intraoperatively, silowing 50 levels (75%) with motion and 17 levels without motion. We reviewed 60 levels by using AP roentgenograms. There were 44 (73%) motion segments intraoperatively and 16 segments without motion. There were 24 (40%) false negative pseudarthroses and 7 (12%) false positive pseudoarthroses, totaling a 57%. preoperative inaccuracy rate.
The fusion rates at each intervertebral level were calculated to ascertain
the frequency of pseudarthrosis (Table 3 ). Of the 35 levels observed
at the L5S1 motion segment in surgery, 21 patients (60%) were found to have
motion greater than 2mm, and 14 patients (40%) were determined to have no
motion or motion less than or equal to 2 mm. Reviewing the L45 motion segment,
25 patients (78%) of the 32 patients were found to have pseudarthrosis and
7 patients (22%) appeared solidly fused. Of the L34 and L23 intervertebral
levels 10 patients (100%) and 1 patient respectively were observed to have
motion greater than 2 mm.
In patients followed up for over 1 year after AILF, it was noted that
three out of three who had previous solid PLF at AILF did not improve at
all. Five of 11 (45%) who had revision by AILF and did not solidify all
their fused segments improved significantly. In those who solidified all
their nonfused segments, 17 of 20 (85%) improved by one or more categories
from the preoperative state using excellent, good, fair and poor clinical
categories. Twelve of 20 (60%) were in the good and excellent categories.
Discussion
According to Chafetz et al. [1] 30%40% of patients will have persistent pain or recurrent pain after fusion. The frequency of patients with pain due to pseudarthrosis has yet to be determined. Frymoyer et al. [2] found a pseudarthrosis rate of 26% in a study of 96 patients with fusions and 36 patients without fusions. No significant difference between the two groups in the postoperative examination could be appreciated.
Cleveland et al. [3] found varying pseudarthrosis rates when reviewing a patient population of 594. The pseudarthrosis rates at the L5S1, L4S1, L3S1 levels were 3%, 17%, and 33%, respectively with the average being 20%. Bending biplanar radiographes were used to define the psuedarthrosis rate. In our study with the use of biplanar x-rays using only the lateral bending films, the clinical pseudarthrosis rate confirmed at surgery combining all levels in our study was 45%. The surgical pseudarthrosis rates at the L5S1, L45 and L34 levels were 60%, 78%, and 100%, respectively. However, our patient population was skewed because only symptomatic patients were reviewed.
The inaccuracy of diagnosis was even higher when evaluating only the
AP lumbar films. The total pseudarthrosis rate was 73%, but the inaccuracy
of diagnosing the pseudarthrosis was 52%, 40% false negative and 12% false
positive. From the study by Cleveland et al. [3] the diagnostic failures were 12% with flat radiographs and clinical judgment. From a study by Dawson at al. [4] an inaccuracy rate of 50% was determined with full spine (AP) radiographs and 82% accuracy rate for AP, lateral and oblique films.
Reports have stated that CT scanning with MPR provided the greatest accuracy in diagnosing pseudarthrosis. Lang et al [5] studied 30 patients with 3D CT scanning and 4 patients with surgical confirmation. Three patients had a pseudarthrosis repair and the fourth patient had pseudarthrosis
identified by surgical identification only. The clearest display of pseudarthrosis is by CT scans using curved coronals rather than just axial CT or 3D CT scans. Laasonen and Soini [7] reviewed 48 patients with pain after fusion by evaluating the patients' CT scans. Of the 27 main lesions detected by CT, 21 were confirmed at surgery and 6 CT scans were incorrect. There was a failure of fusion in 52%, and 16 of the 25 (64%) cases reviewed by plan films and bending films were false negative. However, using anterior exploration of the interspace to assess motion, the CT scans in our study were not as specific as evidenced by a 42% inaccuracy rate with 31% false negative and 11% false positive rates. The total pseudarthrosis rate was 67%.
There is a need to improve the diagnostic capabilities to detect pseudarthrosis; other studies are needed to increase the specificity and sensitivity. Stokes et al. [6] and Pearcy [8] have concluded that biplanar X ray increases the accuracy of diagnosis when enhanced to AP X rays or bending X rays. An even more promising diagnostic tool is stereophotogrammetry which entails placing metallic beads on the posterior elements at the time of surgery [9] . Our own preliminary research indicates lateral flexion/extension biplanar tomography is an accurate test for spinal motion after the fusion procedure.
In conclusion, the accuracy of evaluating FBSS patients for pseudarthrosis was extremely low whether using CT scanning with MPR, bending films or plain X rays, singly or in combination. In view of this, other means to accurately determine the integrity of a fusion mass must be considered. Results after AILF showed most patients with solidification of nonsolid fusions were clinically improved. Those who already had solid fusions or did not consolidate usually did not improve. This implies that small amounts of anterior element motion in abnormal segments heretofore considered by most to be of no clinical signifigance are indeed frequent pain generators in FBSS.
References
1. Chafetz N, Cann CE, Morris JM, Stenbach, US, Goldberg HI (1987) Pseudarthrosis following lumbar fusion: Detection by direct coronal CT scanning. Radiology 162:803805 back
2. Frymoyer JW, Hanley EN, Howe J, Kuhlmann D, Matteri RE (1979) A comparison of radiographic findings in fusion and nonfusiom patients 10 or more years following lumbar disk surgery. Spine 4:435440 back
3. Cleveland M, Bosworth DM, Thompson FR (1948) Pseudarthrosis in the lumbosacral spine. J Bone Joint Surg [Am] 30:13021312 back
4. Dawson EG, Clader TJ, Bassett LW (1985) A comparison of different methods used to diagnose pseudarthrosis following posterior spinal fusion for scoliosis. J Bone Joint Surg [Am] 67:11531159 back
5. Lang P, Genant HK, Chafetz N, Steiger P, Morris JM (1988) Threedimensional computer tomography and multiplanar reformations in the assessment of pseudarthrosis in posterior lumbar fusion patients. Spine 13:6975 back
6. Stokes IAF, Wilder DG, Frymoyer JW, Pope MH (198l) Assessment of patients with low back pain by biplanar radiographic measurement of intervertebral motion. Spine 6:233240 back
7. Laasonen EM, Soini J (1989) Lowback pain after lumbar fusion: Surgical computer tomographic analysis. Spine 14:210213 back
8. Pearcy M (1982) Assessment of bony union after interbody fusion of the lumbar spine using a biplanar radiographic technique. J Bone Joint Surg [Br] 64:225232 back
9. Morris J, Chafetz N, Baumrind S, Genant H, Korn EL (1985) Stereophotogrammetry of the lumbar spine: A technique for the detection of pseudarthrosis. Spine 10:368375 back |
