Content uploaded by Takashi Kaito
Author content
All content in this area was uploaded by Takashi Kaito on Feb 10, 2015
Content may be subject to copyright.
J Neurosurg: Spine / Volume 12 / June 2010 671
Wh e n instability of the lumbar spine causes per-
sistent pain or neurological impairment, it is
usually treated using spinal fusion. Among the
various fusion techniques, PLIF with instrumentation
is one option. The technique has the advantage of both
360° decompression and fusion, and it has provided sat-
isfactory clinical results and high fusion rates.21,28,33 Un-
fortunately, spinal fusion alters the normal biomechanics
of the spine, and the loss of motion at the fused level is
compensated by increased motion and load at the other
unfused segments. Possibly as a result of altered biome-
chanics at the adjacent mobile segments, ASD, a seri-
ous complication of PLIF, can occur. The actual cause
of ASD, however, remains unknown. Some authors
have attributed ASD to the altered biomechanical prop-
erty due to immobilization of the fusion segment,4,19,20,23
whereas others have argued that ASD is nothing more
than a common aging process rather than a consequence
of spinal fusion.10,13,17,18,26,34
Several risk factors for the development of ASD have
been proposed to date, such as the number of segments
involved, sagittal and/or coronal imbalance, stiffness
of fusion, laminar inclination, and facet tropism.6,22,24,27
Most of these factors are unavoidable during spinal fu-
sion and are beyond the surgeon’s control. In this study,
we focus our attention on the degree of disc height dis-
traction during surgery as a potential causative factor
in the development of ASD. The distraction of the disc
height by cage or bone insertion could affect adjacent
disc and facet joints, inducing early ASD.
Induction of early degeneration of the adjacent segment
after posterior lumbar interbody fusion by excessive
distraction of lumbar disc space
Clinical article
Ta k a s h i ka i T o , M.D.,1 No b o r u ho s o N o , M.D.,2 Yo s h i h i r o Mu k a i , M.D.,2
Ta k a h i r o Ma k i N o , M.D.,1 Ta k e s h i Fu j i , M.D.,2 a N D ka z u o Yo N e N o b u , M.D.1
1Department of Orthopedics, National Hospital Organization Osaka Minami Medical Center,
Kawachinagano; and 2Department of Orthopedics, Osaka Kosei-nenkin Hospital, Fukushima-ku,
Osaka, Japan
Object. Spinal fusion at the L4–5 disc space alters the normal biomechanics of the spine, and the loss of mo-
tion at the fused level is compensated by increased motion and load at the other unfused segments. This may lead to
deterioration of the adjacent segments of the lumbar spine, called adjacent-segment disease (ASD). In this study, the
authors investigate the distracted disc height of the fused segment, caused by cage or bone insertion during surgery,
as a novel risk factor for ASD after posterior lumbar interbody fusion (PLIF).
Methods. Radiographic L3–4 ASD is dened by development of spondylolisthesis greater than 3 mm, a decrease
in disc height of more than 3 mm, or intervertebral angle at exion smaller than −5°. Symptomatic ASD is dened
by a decrease of 4 points or more on the Japanese Orthopaedic Association scale. Eighty-ve patients with L-4
spondylolisthesis treated by L4–5 PLIF underwent follow-up for more than 2 years (mean 38.8 ± 17.1 months). The
patients were divided into 3 groups according to the nal outcome. Group A comprised those patients without ASD
(58), Group B patients had radiographic ASD (14), and Group C patients had symptomatic ASD (13).
Results. The L4–5 disc space distraction by cage insertion was 3.1 mm in the group without ASD, 4.4 mm in
the group with radiographic ASD, and 6.2 mm in the group with symptomatic ASD, as measured using lateral spinal
radiographs just after surgery. Multivariate analysis showed that distraction was the most signicant risk factor.
Conclusions. The excessive distraction of the L4–5 disc space during PLIF surgery is a signicant and poten-
tially avoidable risk factor for the development of radiographic, symptomatic ASD.
(DOI: 10.3171/2009.12.SPINE08823)
ke Y Wo r D s • disc height • adjacent segment disease • complication
Abbreviations used in this paper: ASD = adjacent-segment dis-
ease; JOA = Japanese Orthopaedic Association; PLIF = posterior
lumbar interbody fusion; ROM = range of motion.
J Neurosurg Spine 12:671–679, 2010
T. Kaito et al.
672 J Neurosurg: Spine / Volume 12 / June 2010
Methods
Between 1999 and 2003, 97 consecutive patients
at our institution underwent L4–5 PLIF using the same
pedicle screw system and intervertebral cages (Steffee
VSP and Brantigan I/F Cage, Depuy Spine, Inc.) for L-4
spondylolisthesis by 5 surgeons. Eighty-ve patients (29
men and 56 women) who had complete medical records
including radiographs and a minimum of 2 years of post-
operative follow-up were included in this retrospective
study. The mean age at surgery was 64.1 years (range
36–83 years), and the average follow-up period was 38.8
months (range 24–84 months). No patient included in the
study had decompression or fusion procedures other than
L4–5 PLIF. The number of operations performed by each
surgeon was as follows: 54 by Surgeon A, 18 by Surgeon
B, 8 by Surgeon C, 3 by Surgeon D, and 2 by Surgeon
E. Neurological status was assessed before surgery, at a
maximally recovered time during the follow-up, and at
the nal visit using a scoring system proposed by the JOA
scale (29-point system).11
The achievement of radiographic fusion was deter-
mined by the presence of continuous trabecular bone
bridging across the disc space and the absence of screw
loosening and residual motion at the fused segment on
exion and extension lateral radiographs. We made
measurements on radiographs and CT scans using im-
age measuring software Image-J (National Institutes of
Health) after digitizing the lms (SIERRA plus, VIDAR
System Co.).
Measurements Taken Using the Plain Radiograph
The L4–5 vertebral height (H) was dened as the
distance between the midpoint of the upper endplate of
the L-4 and lower endplate of the L-5 vertebrae on the
lateral radiograph in neutral position (Fig. 1A). The L4–5
vertebral height was measured before surgery, just after
surgery, and at the nal visit (or just before surgery at the
L3–4 level to treat ASD).9 The parameters related to the
L4–5 vertebral height were dened as follows: ∆Hmax =
(H just after surgery)−(H before surgery) and ∆Hnal = (H
at nal visit)−(H before surgery).
The following additional measurements were taken
from the plain radiographs at L3–4 before surgery and at
the nal visit and at L4–5 before surgery: 1) anterolisthe-
sis at exion (a, in mm); 2) retrolisthesis at extension (b,
in mm); 3) distance of translation (a + b, in mm); 4) inter-
vertebral angle at exion (c, in degrees [the angle is made
by the endplates of the disc space; lordosis is a positive
value); 5) intervertebral angle at extension (d, in degrees);
6) ROM (d−c, in degrees); 7) disc height (in mm); 8) L-3
laminar inclination (see Fig. 1B);16 9) L4–5 fusion angle
just after surgery and again at the nal visit (Fig. 1C); and
10) lumbar lordosis between L-1 and S-1 before surgery
and at the nal visit (in degrees, the angle made by the
upper endplate of L-1 and S-1 in neutral position).
Measurements on CT Scans
The right and left facet angles (γ1 and γ2) were mea-
sured at the L3–4 disc level on preoperative CT images,
and then the sum of the right and left facet angles (γ1
+ γ2) was dened as facet sagittalization. The difference
between the right and left facet angle (γ1−γ2) was dened
as facet tropism (Fig. 1D).29 The degree of L3–4 facet
joint degeneration was classied into Grades 0–3 by the
Weishaupt grading system, as follows. Grade 0, normal
facet joint space (2–4 mm width); Grade 1, narrowing of
the facet joint (< 2 mm) and/or small osteophytes and/
or mild hypertrophy of the articular process; Grade 2,
narrowing of the facet joint space and/or moderate osteo-
phytes and/or moderate hypertrophy of the articular pro-
cess and/or mild subarticular bone erosions; and Grade
3, narrowing of the facet joint space and/or large osteo-
phytes and/or severe hypertrophy of the articular process
and/or severe subarticular bone erosion and/or subchon-
dral cysts.30
The radiographic ASD at L3–4 was dened using
plain radiographs taken before surgery and at the nal
visit, irrespective of the presence or absence of concomi-
tant clinical symptoms. Radiographic ASD comprised ei-
ther development of L3 antero- or retrolisthesis more than
3 mm, a decrease in L3–4 disc height of more than 3 mm,
or intervertebral angle at exion smaller than −5°.20 Clini-
cal deterioration by L3–4 ASD was dened as a decrease
by 4 points or more on the JOA scale accompanied by
neurological impairment in accordance with L3–4 canal
stenosis based on MR imaging, which was assessed every
6 months.
The adjacent segment L5–S1 was not investigated in
this study since the degeneration at L5–S1 is frequently
found preoperatively and rarely causes clinical symp-
toms.1,4,17,20,21 Therefore, no patient in our study received
surgery at the L5–S1 level.
Patients were divided into 3 groups by clinical and
radiographic status at the time of the nal visit. Group
A comprised those patients who had neither radiographic
ASD nor clinical deterioration (as dened above). Group
B patients had radiographic ASD without clinical dete-
rioration (radiographic ASD group). Group C patients had
clinical deterioration caused by spinal stenosis at L3–4
with or without radiographic ASD (symptomatic ASD
group, including the patients who underwent surgery for
L3–4 ASD).
Statistical Analysis
A paired t-test, Fisher exact test, chi-square test, and
Mann-Whitney U-test were used to compare values. Mul-
tivariate logistic regression was performed using vari-
ables with p < 0.20 in univariate analysis. A p value <
0.05 was considered statistically signicant.
Results
Using the above criteria, 58 patients (17 men and 41
women) were classied into Group A, 14 patients (6 men
and 8 women) into Group B, and 13 patients (7 men and
6 women) into Group C. In Group C, 10 of 13 patients
fullled the denition of radiographic ASD, and 11 of 13
patients had undergone surgery for L3–4 ASD.
These 3 groups were comparable in age, sex ratio,
and JOA score before surgery and postoperatively (sex:
p > 0.1, Fisher exact test; other items: p > 0.3, Mann-Whit-
J Neurosurg: Spine / Volume 12 / June 2010
A novel risk factor for adjacent-segment disease after PLIF
673
ney U-test, as shown in Table 1). In all cases, L4–5 fusion
was successfully achieved at the nal visit in all patients
studied; therefore, the fusion rate at L4–5 was 100%.
Distracted Distance of L4–5 Disc Space by Cage Insertion
on Plain Radiography
The distracted distance of L4–5 disc space by cage
insertion, or mean ∆Hmax, was 3.1 ± 2.2 mm in the group
without ASD, 4.4 ± 1.7 mm in the group with the L3–4
radiographic ASD, and 6.2 ± 2.6 mm in the group with
the symptomatic ASD. All measurements are described
as mean ± SD. The ∆Hmax in Groups B and C was signi-
cantly greater than that in Group A (p = 0.04 [A vs B],
p = 0.0005 [A vs C], p = 0.07 [B vs C]; Mann-Whitney
U-test).
The mean ∆Hnal was 1.3 ± 2.7 for Group A, 1.5 ±
2.4 for Group B, and 4.3 ± 2.3 for Group C. The ∆Hnal of
Group C was signicantly greater than that of Groups A
and B (p = 0.9 [A vs B], p = 0.0004 [A vs C], p = 0.009 [B
vs C]; Mann-Whitney U-test) (Fig. 2).
Other Radiographic Results Before and Just After Surgery
Based on Plain Radiographs Before Surgery and Just After
Surgery
There were no signicant differences among the 3
groups regarding L-3 anterolisthesis in exion, L-3 ret-
Fig . 1. Methods of measurement of the various parameters. A: The L4–5 vertebral height (H) was measured from the mid-
point of the upper endplate of L-4 to the midpoint of lower endplate of the L-5 vertebra. B: The L-3 laminar inclination (α) was
defined as the angle formed by a straight line connecting the base of the superior facet joint and the base of the inferior facet
joint and a horizontal line bisecting the vertebral body. C: The L4–5 fusion angle (β) was an angle made by tangent lines of the
upper endplate of L-4 upper and the lower endplate of L-5. D: Facet sagittalization and tropism were defined as the sum of the
right and left facet angle (γ1 + γ2) and the difference between the right and left facet angle (γ1–γ2), respectively.
TABLE 1: Characteristics of each group*
Variable Total Group A Group B Group C p Value
no. of patients 85 58 14 13
no. male/female 29:56 17:41 6:8 7:6 NS
mean age (yrs) 64.1 ± 8.6 63.4 ± 8.6 65.5 ± 9.2 66.0 ± 8.3 NS
mean preop JOA score 15.4 ± 4.0 15.1 ± 4.2 15.7 ± 3.3 16.4 ± 3.8 NS
me an JOA score at best time
(preop for ASD)
24.6 ± 4.1 25.6 ± 3.3 26.5 ± 2.2 25.6 ± 2.2
(18.5 ± 3.5)
NS
mean follow-up period (mos) 38.8 ± 17.1 38.6 ± 16.7 41.6 ± 16.6 36.7 ± 20.6 NS
* Mean values are represented as ± SDs. Abbreviation: NS = not signicant.
T. Kaito et al.
674 J Neurosurg: Spine / Volume 12 / June 2010
rolisthesis in extension, distance of L-3 translation, L3–4
intervertebral angle at exion and extension, L3–4 ROM,
L-3 laminar inclination, L3–4 disc height, L-4 anterolis-
thesis at exion, L-4 retrolisthesis at extension, and L4–5
intervertebral angle at exion and extension (Mann-Whit-
ney U-test) (Table 2).
The preoperative L4–5 disc height was 8.0 ± 2.0 mm
in Group A, 6.6 ± 2.5 mm in Group B, and 5.5 ± 2.0
mm in Group C, with the value in Group C signicant-
ly smaller than that in Group A (p = 0.09 [A vs B], p =
0.0004 [A vs C], p = 0.4 [B vs C]; Mann-Whitney U-test).
The distance of L-4 translation in Groups B and C was
smaller than that in Group A (p = 0.02 [A vs B], p = 0.08
[A vs C]; Mann-Whitney U-test).
The L4–5 ROM in Group A was greater than that
in Groups B and C, although the difference was not sta-
tistically signicant (p = 0.06 [A vs B], p = 0.1 [A vs C],
p = 0.7 [B vs C]; Mann-Whitney U-test). Lumbar lordosis
before surgery and the L4–5 fusion angle just after sur-
gery were not signicantly different among the 3 groups
(Mann-Whitney U-test; Tables 2 and 3).
Measurements Based on Plain Radiographs at the Final
Visit
The postoperative decrease in L3–4 disc height mea-
sured on the most recent radiograph was 0.9 ± 0.7 mm in
Group A, 2.2 ± 1.8 mm in Group B, and 2.2 ± 1.8 mm in
Group C. The decrease in L3–4 disc height in Groups B
and C was signicantly greater than that in Group A (p =
0.03 [A vs B], p = 0.02 [A vs C], p > 0.9 [B vs C]; Mann-
Whitney U-test).
The L-3 anterolisthesis at exion was 0.6 ± 1.1 mm
in Group A, 2.8 ± 2.5 mm in Group B, and 2.2 ± 3.2 mm
in Group C. The degree of L-3 anterolisthesis in Groups
B and C was signicantly greater than that in Group A
(p = 0.0003 [A vs B], p = 0.05 [A vs C], p = 0.3 [B vs C];
Mann-Whitney U-test). The L-3 retrolisthesis at extension
was 1.6 ± 1.7 mm in Group A, 1.5 ± 1.9 mm in Group B,
and 3.5 ± 3.1 mm in Group C. The degree of retrolisthesis
at extension in Group C was greater than that in Groups
A and B, although it was not signicant (p = 0.7 [A vs
B], p = 0.1 [A vs C], p = 0.2 [B vs C]; Mann-Whitney
U-test). The distance of L-3 translation was 1.7 ± 1.3 mm
in Group A, 2.5 ± 1.4 mm in Group B, and 3.8 ± 2.2 mm
in Group C. The L-3 translation value in Groups B and C
was signicantly greater than that in Group A (p = 0.04
[A vs B], p = 0.003 [A vs C]; Mann-Whitney U-test). The
L-3 translation value in Group C was greater than that in
Group B, although it was not signicant (p = 0.2 [B vs C],
Mann-Whitney U-test).
The L3– 4 intervertebral angles at exion and exten-
sion were not signicantly different among the groups
(Mann-Whitney U-test). The L3–4 ROM was 6.0 ± 3.5°
in Group A, 6.3 ± 5.3° in Group B, and 9.2 ± 3.8° in
Group C. The value of ROM for Group C was signi-
cantly greater than that of Group A (p = 0.8 [A vs B], p =
0.08 [B vs C], p = 0.009 [A vs C]; Mann-Whitney U-test)
Lumbar lordosis and the L4–5 fusion angle at the nal
visit were not signicantly different among the 3 groups
(Mann-Whitney U-test) (Table 3).
Computed Tomography Data
Regarding the CT data, facet sagittalization and
facet tropism before surgery were not signicantly dif-
ferent among the groups (Mann-Whitney U-test) (Table
2). The preoperative L3–4 facet joint degeneration grade,
based on the CT scans in each group, is shown in Table
4. The degeneration grade was not signicantly different
between groups (p = 0.34 [A vs B], p = 0.11 [A vs C], p =
0.51 [B vs C]; Mann-Whitney U-test [p = 0.37, chi-square
test]).
Fig . 2. Graph showing that ∆Hmax is significantly greater in Groups B and C compared with Group A, and ∆Hfinal is significantly
greater in Group C compared with Groups A and B (Mann-Whitney U-test).
J Neurosurg: Spine / Volume 12 / June 2010
A novel risk factor for adjacent-segment disease after PLIF
675
Incidence of ASD for Each Surgeon
To evaluate the surgeon’s factor, the incidence of
ASD for each surgeon was investigated. The incidence
was not signicantly different among surgeons (p = 0.46,
chi-square test) (Table 5).
Multiple Logistic Regression Analysis
To compare the relative impact of these variables for
ASD, multiple logistic regression analysis was performed.
The result of univariate analysis between the control group
(Group A) and the ASD group (Groups B and C) is shown
in Table 6. Variables with p < 0.20 in univariate analysis
(before surgery: L3–4 intervertebral angle [exion], L3–4
disc height, L-4 retrolisthesis [extension], L-4 distance of
translation, L4–5 intervertebral angle [extension], L4–5
disc height, lumbar lordosis at L1–S1, and L3–4 facet
joint degeneration; immediately after surgery: ∆Hmax; and
at the nal visit: ∆Hnal and lumbar lordosis at L5–S1) and
a variable reported as a risk factor for ASD in previous
reports (L4–5 fusion angle at nal visit) were analyzed as
dependent variables by the forced entry method. Analysis
showed that ∆Hmax was a sole signicant independent risk
factor of ASD after PLIF (Table 7).
Surgery for ASD
Eleven of 13 patients in Group C underwent addi-
tional surgery to treat ASD at L3– 4. The mean period
between the index surgery and the additional surgery was
29. 9 ± 18.3 months (range 9–65 months), and the JOA
score was reduced to 16.7 ± 2.0 (range 6–24) just before
surgery for ASD. The operative procedures for the ASD
were PLIF in 4 patients, partial laminectomy in 6, and
posterolateral fusion in 1. The JOA score signicantly im-
proved following these additional surgeries for ASD.
Discussion
Several studies have argued that ASD is nothing
TABLE 2: Imaging results before surgery*
Variable Total Group A Group B Group C p Value
plain radiography
L-3 & L3–4
anterolisthesis (ex, mm) 0.4 ± 0.9 0.3 ± 0.9 0.4 ± 1.1 0.6 ± 1.1 NS
retrolisthesis (ext, mm) 0.8 ± 1.2 0.8 ± 1.1 0.3 ± 0.7 1.6 ± 1.7 NS
distance of translation (mm) 0.9 ± 1.1 0.8 ± 1.0 0.4 ± 1.2 1.5 ± 1.1 NS
intervertebral angle (ex, °) 1.6 ± 3.2 1.7 ± 3.0 0.3 ± 3.2 1.9 ± 4.1 NS
intervertebral angle (ext, °) 8.3 ± 3.4 8.7 ± 3.3† 6.0 ± 3.1 9.5 ± 3.4‡
ROM (°) 6.8 ± 3.2 6.9 ± 3.1 5.6 ± 3.4 7.6 ± 3.5 NS
disc height (mm) 9.8 ± 1.8 10.0 ± 1.8 9.3 ± 1.4 9.5 ± 2.0 NS
laminar inclination (°) 123.6 ± 4.1 123.5 ± 3.9 123.3 ± 4.4 124.4 ± 5.1 NS
L-4 & L4–5
anterolisthesis (ex, mm) 6.6 ± 2.7 6.4 ± 2.7 7.4 ± 3.3 6.9 ± 2.3 NS
retrolisthesis (ext, mm) −4.3 ± 2.9 −3.8 ± 2.5 −5.7 ± 3.6 −5.2 ± 3.3 NS
distance of translation (mm) 2.3 ± 1.6 2.6 ± 1.5 1.7 ± 1.8§ 1.7 ± 1.6
intervertebral angle (ex, °)¶ −3.5 ± 4.9 −3.6 ± 4.5 −4.4 ± 6.8 −2.0 ± 4.5 NS
intervertebral angle (ext, °) 6.6 ± 4.1 7.1 ± 3.9 4.1 ± 5.2 6.8 ± 2.9 NS
ROM (°) 10.0 ± 3.8 10.6 ± 3.7 8.6 ± 4.1 8.8 ± 3.4 NS
disc height (mm) 7.4 ± 2.3 8.0 ± 2.0 6.6 ± 2.5 5.5 ± 2.0**
lumbar lordosis at L1–S1 (°) 35.4 ± 11.4 36.5 ± 10.9 33.3 ± 13.9 33.4 ± 10.3 NS
CT scanning
L3–4
facet sagittalization (°) 63.8 ± 17.0 63.8 ± 17.0 61.3 ± 19.5 63.0 ± 11.8 NS
facet tropism (°) 5.3 ± 4.6 4.7 ± 4.0 7.2 ± 6.7 5.8 ± 4.1 NS
facet joint degeneration†† 2.5 ± 0.6 2.6 ± 0.6 2.4 ± 0.5 2.2 ± 0.7 NS
* Values are expressed as the means ± SDs. Abbreviations: ext = extension; ex = exion.
† p < 0.01 vs Group B.
‡ p < 0.05 vs Group B.
§ p < 0.05 vs Group A.
¶ The L3–4 intervertebral angle at extension was signicantly smaller than that of other groups, but the L3–4 ROM was not
signicantly different among groups.
** p < 0.01 vs Group A.
†† Values for facet joint degeneration are Weishaupt grade.
T. Kaito et al.
676 J Neurosurg: Spine / Volume 12 / June 2010
more than the normal degenerative process rather than a
consequence of spinal fusion.10,13,17,18,26,34 In many of these
reports, however, the follow-up period is too long to min-
imize the inuence of age-related degeneration (mean
follow-up periods of 13,13 15, 26 and 22 years10). In addi-
tion, the inconsistent denition of ASD and patient char-
acteristics prevent analysis of risk factors for ASD. In our
study, the mean follow-up period was only 3 years, which,
we believe, is short enough to exclude the inuence of the
aging process at the adjacent segment of the PLIF.
A PLIF theoretically provides 360° decompression
and fusion of the affected spinal segment and also releas-
es entrapped nerve roots by distracting the disc space and
neural foramen.32 In the original PLIF procedure, iliac
bones alone were grafted in the disc space. It was, how-
ever, difcult to maintain the disc height after surgery
due to the collapse of the grafted iliac bone. Delayed or
nonunion of PLIF often occurred. Even if the bony union
was successful, the segment was often fused in a state of
so-called “collapsed union,” which is the gradual reduc-
tion of disc height over the postoperative period.7,31 The
introduction of the carbon cage combined with posterior
instrumentation solved this problem. It ensured the main-
tenance of disc height until the completion of bony fu-
sion3, 28 and improved fusion rate dramatically.3,15,19,21, 28, 33
Therefore, the use of these devices has become standard
procedure in PLIF.
In patients with severely reduced disc height, how-
ever, cages cannot be inserted into the disc space without
using forced distraction. Our assumption is that the dis-
traction of the L4–5 disc space by cage insertion exerts
signicant mechanical stress on the adjacent L3–4 seg-
ment. In the present study, the distracted distance of the
L4–5 disc space by cage insertion, or mean ∆Hmax, was
3.1 mm in Group A and was not associated with either
radiographic or symptomatic ASD. In Groups B and C,
however, the ∆Hmax values were 4.4 and 6.2 mm, respec-
tively, and this increase in ∆Hmax was accompanied by
signicant ASD. The multivariate analysis showed ∆Hmax
was the sole signicant independent risk factor of ASD
after PLIF among the various factors analyzed. These re-
sults indicate that distraction of the disc space in PLIF
beyond a certain amount could induce ASD, and we spec-
ulate that the cause of the ASD is altered biomechanics.
If the L4–5 disc space is lifted up via cage insertion,
the increase in distance does not result in an increase in
the total height of the spinal column but is partially ab-
sorbed somewhere in the spine. It is natural to consider
the adjacent disc and facet joints as the most suitable tis-
sues for absorption of this increase in distance. The com-
pression force on the L3–4 disc and facet joints could
result in ASD (Fig. 3). Of course, this hypothesis, which
assumes an alteration in the biomechanics of the spine as
TABLE 3: Radiographic results immediately following surgery and at nal visit
Variable Total Group A Group B Group C
immediately postop
L4–5 fusion angle (°) 10.5 ± 4.6 10.7 ± 4.4 10.0 ± 3.6 10.3 ± 6.5 NS
at nal visit
decrease of L3–4 disc height (mm) 1.3 ± 1.3 0.9 ± 0.7 2.2 ± 1.8* 2.2 ± 1.8*
L-3 anterolisthesis (ex, mm) 1.2 ± 2.0 0.6 ± 1.1 2.8 ± 2.5† 2.2 ± 3.2*
L-3 retrolisthesis (ext, mm) 1.8 ± 2.1 1.6 ± 1.7 1.5 ± 1.9 3.5 ± 3.1 NS
distance of L-3 translation (mm) 2.1 ± 1.7 1.7 ± 1.3 2.5 ± 1.4* 3.8 ± 2.2†
L3–4 interver tebral angle (ex, °) 2.7 ± 4.3 3.5 ± 3.8 2.1 ± 4.2 0.2 ± 5.5 NS
L3–4 interver tebral angle (ext, °) 9.3 ± 4.0 9.5 ± 3.8 8.3 ± 3.6 9.4 ± 5.3 NS
L3–4 ROM (°) 6.6 ± 4.0 6.0 ± 3.5 6.3 ± 5.3 9.2 ± 3.8†
L4–5 fusion angle (°) 10.0 ± 5.1 11.7 ± 6.1 9.2 ± 3.3 11.7 ± 6.1 NS
lumbar lordosis at L1–S1 (°) 34.4 ± 20.8 35.4 ± 11.1 32.4 ± 10.1 32.3 ± 10.3 NS
* p < 0.05 vs Group A.
† p < 0.01 vs Group A.
TABLE 4: The L3– 4 facet joint degeneration grade based on
CT scanning ndings
No. of Patients
Group Grade 0 Grade 1 Grade 2 Grade 3
A 2 22 33 1
B 0 8 6 0
C 2 6 5 0
TABLE 5: The number of cases of ASD per surgeon
No. of Patients
Group
Surgeon
A
Surgeon
B
Surgeon
C
Surgeon
D
Surgeon
E
A34 14 721
B12 1100
C 83011
total 54 18 832
J Neurosurg: Spine / Volume 12 / June 2010
A novel risk factor for adjacent-segment disease after PLIF
677
an explanation of our data, has not yet been proven in a
biomechanical laboratory.
Until now, the distraction of disc space at spinal fu-
sion has not been considered a possible cause of ASD.
Only 1 study, which used a nite element model to evalu-
ate facet joint stress at the adjacent level (L2–3), found
an 8% increase in stress due to a 2-mm distraction of the
fused segment (L3–4).24 Nevertheless, some studies seem
to indirectly support our hypothesis. For example, PLIF
for disc herniation has been reported to yield a lower in-
cidence of ASD than PLIF for spondylolisthesis.19 This
nding may be due to the fact that the preoperative disc
height in herniation cases is, in general, larger than that
of spondylolisthesis and so the distraction by cage inser-
tion may be relatively smaller in disc herniation cases.
In addition, the incidence of ASD after instrumented
posterolateral fusion, where distraction is rarely applied
and requires no insertion of cages, is reported to be lower
than that after instrumented PLIF.1,2,12,25 Finally, anterior
lumbar fusion with iliac graft, in which the distracted
disc space often returns to the original disc height (col-
lapsed union) before bony fusion, yields a low incidence
of ASD.5,8 All of these reports, although they do not spec-
ify the distracted distance and intraoperative soft-tissue
and facet joint injury or alignment of fused segments, are
different because of inhomogeneous disorders or surgical
approaches, seem to indirectly support our hypothesis.
Conicting results have been reported with respect
to the relationship between radiographic ASD and symp-
tomatic ASD because of varying denitions of ASD.1,4,15
In our study, radiographic L3–4 ASD was dened by
the development of antero- or retrolisthesis greater than
3 mm, the decrease of disc height more than 3 mm, or
intervertebral angle at exion less than −5 °. 20 Of course,
patients with neurological symptoms did not necessarily
fulll these criteria of radiographic ASD. Indeed, 23%
(3 of 13) of the patients who had symptomatic ASD did
not meet our denition of radiographic ASD. Conversely,
42% (10 of 24) of patients who had radiographic evidence
of ASD were symptomatic. These results provide 81% (10
of 13) sensitivity and 77% (58 of 72) specicity, which
makes our denition of radiographic ASD reasonable
enough to predict the development of neurological symp-
toms.
The L4–5 fusion angle (Cobb angle) and lumbar
lordosis are reported to be related to the development of
ASD.4,14 However, they were not identied as risk factors
for ASD in our study. We believe that the acquisition of
lumbar lordosis at the fused segment (mean L4–5 fusion
angle was approximately 10° in all groups) by using pedi-
TABLE 6: Univariate analysis between the control group and the
ASD group
Variable
Control
Group
(Group A)
ASD Group
(Groups
B & C)
p
Value*
preop
L-3 & L3–4
anterolisthesis (ex, °) 0.3 ± 0.9 0.3 ± 1.4 0.31
retrolisthesis (ext, °) 0.8 ± 1.1 0.9 ± 1.4 0.90
distance of translation (mm) 0.8 ± 1.0 1.1 ± 1.1 0.48
intervertebral angle (ex, °) 1.7 ± 3.0 1.1 ± 4.9 0.12
intervertebral angle (ext, °) 8.7 ± 3.3 8.8 ± 4.4 0.57
ROM (°) 6.9 ± 3.1 7.7 ± 4.8 0.87
disc height (mm) 10.0 ± 1.8 9.4 ± 1.7 0.07
laminar inclination (°) 123.5 ± 3.9 123.8 ± 4.7 0.71
facet sagittalization (°) 63.8 ± 17.0 62.1 ± 16.0 0.57
facet tropism (°) 4.7 ± 4.0 6.5 ± 5.6 0.22
facet joint degeneration 2.5 ± 0.6 2.3 ± 6.2 0.10
L-4 & L4–5
anterolisthesis (ex, °) 6.4 ± 2.7 7.1 ± 2.8 0.23
retrolisthesis (ext, °) −3.8 ± 2.5 −5.4 ± 3.4 0.04
distance of translation (mm) 2.6 ± 1.5 1.8 ± 1.6 0.01
intervertebral angle (ex, °) −3.6 ± 4.5 −3.0 ± 6.0 0.57
intervertebral angle (ext, °) 7.1 ± 3.9 4.4 ± 4.4 0.19
ROM (°) 10.6 ± 3.7 8.4 ± 4.0 0.02
disc height (mm) 8.0 ± 2.0 6.1 ± 2.2 0.01
lumbar lordosis at L1–S1(°) 36.5 ± 10.9 33.3 ± 12.1 0.11
immediately postop
ΔH
max (mm) 4.4 ± 1.7 5.3 ± 2.3 0.00
L4–5 fusion angle (°) 10.7 ± 4.4 10.2 ± 5.1 0.63
at nal visit
L-4 & L4–5
ΔH
nal (mm) 1.3 ± 2.7 2.9 ± 2.7 0.02
L4–5 fusion angle (°) 11.7 ± 6.1 10.4 ± 4.9 0.74
lumbar lordosis at L1–S1 (°) 35.4 ± 11.1 32.0 ± 12.3 0 .12
* Mann-Whitney U-test.
TABLE 7: Multivariate logistic regression for ASD after PLIF
Time Point Risk Factor p Value OR (95% CI)
preop L-3 intervertebral angle
(ex)
0.19 0.82 (0.61–1.10)
L3–4 disc height 0.76 1.77 (0.05–65.52)
L-4 retrolisthesis 0.81 1.53 (0.05– 45.01)
L-4 distance of transla-
tion
0.26 0.07 (0.00–7.50)
L-4 intervertebral angle
(ext)
0.66 1.06 (0.83–1.34)
L4–5 ROM 0.13 0.85 (0.69 –1.05)
L4–5 disc height 0.1 0 0.04 (0.00–1.80)
lumbar lordosis at
L1– S1
0.78 1.01 (0.92–1.12)
facet joint degeneration 0.78 0.84 (0.25–2.83)
immediately
postop
ΔHmax 0.04 57.26 (1.11–2966.52)
at nal visit ΔHnal 0.54 0.39 (0.02–8.27)
lumbar lordosis at
L1– S1
0.46 0.96 (0.86–1.07)
L4–5 fusion angle 0.28 1.09 (0.93–1.27)
T. Kaito et al.
678 J Neurosurg: Spine / Volume 12 / June 2010
cle screws and plate system (Steffee VSP) may have pre-
vented the inuence of malalignment of fused segments
on the development of ASD.
Laminar inclination and facet tropism at the level
adjacent to PLIF have been reported as possible risk fac-
tors for the development of postoperative instability, since
the former could allow anteroposterior translation and
the latter could make the spinal segment susceptible to
asymmetrical rotational instability.22 In our study, how-
ever, they were not identied as signicant risk factors
for ASD. One of the reasons is the difculty in identify-
ing the measuring points on images of degenerated and
often hypertrophied vertebrae and facet joints. Further-
more, these risk factors reect the patient’s anatomical
characteristics, which cannot be controlled by operative
procedure. In contrast, the excessive distraction of the
disc space can be avoided by inserting smaller cages in
PLIF or by applying fusion techniques other than PLIF.
Surgeons, therefore, need to avoid excessive disc space
distraction when performing PLIF.
Conclusions
We investigated risk factors of L3–4 ASD in 85
consecutive patients treated with L4–5 PLIF at an av-
erage follow-up of 3.2 years. The distracted distance of
L4–5 disc space by cage insertion was 3.1 ± 2.2 mm in
the group without ASD, 4.4 ± 1.7 mm in the group with
L3–4 radiographic ASD, and 6.2 ± 2.6 mm in the group
with symptomatic ASD. In addition, the distracted dis-
tance between the ASD groups and the control group was
statistically signicant according to univariate and mul-
tivariate analyses. The excessive distraction of disc space
is, therefore, suggested as a novel risk factor for the devel-
opment of the radiographic and symptomatic ASD.
Disclosure
The authors report no conflict of interest concerning the materi-
als or methods used in this study or findings specified in this paper.
Author contributions to the study and manuscript preparation
include the following. Conception and design: T Kaito, N Hosono.
Acquisition of data: T Kaito, T Makino. Analysis and interpretation
of data: T Kaito. Drafting the article: T Kaito. Critically revising
the article: Y Mukai. Reviewed final version of the manuscript and
approved it for submission: T Kaito, N Hosono, Y Mukai, T Makino,
T Fuji, K Yonenobu. Statistical analysis: T Kaito. Study supervision:
N Hosono, T Fuji, K Yonenobu.
References
1. Aota Y, Kumano K, Hirabayashi S: Postfusion instability at
the adjacent segments after rigid pedicle screw xation for de-
generative lumbar spinal disorders. J Spinal Disord 8:464–
473, 1995
2. Axelsson P, Johnsson R, Strömqvist B: Adjacent segment hy-
permobility after lumbar spine fusion: no association with
progressive degeneration of the segment 5 years after surgery.
Acta Orthop 78:834–839, 2007
3. Brantigan JW, Steffee AD: A carbon ber implant to aid in-
terbody lumbar fusion. Two-year clinical results in the rst 26
patients. Spine 18:2106–2107, 1993
4. Cheh G, Bridwell KH, Lenke LG, Buchowski JM, Daubs MD,
Kim Y, et al: Adjacent segment disease following lumbar/
thoracolumbar fusion with pedicle screw instrumentation: a
minimum 5-year follow-up. Spine 32:2253–2257, 2007
5. Cheung KMC, Zhang YG, Lu DS, Luk KDK, Leong JCY: Re-
duction of disc space distraction after anterior lumbar inter-
body fusion with autologous iliac crest graft. Spine 28:1385–
1389, 2003
6. Chosa E, Goto K, Totoribe K, Tajima N: Analysis of the effect
of lumbar spine fusion on the superior adjacent intervertebral
disk in the presence of disk degeneration, using the three-di-
mensional nite element method. J Spinal Disord Tech 17:
134–139, 2004
7. Cloward RB: Posterior lumbar interbody fusion updated. Clin
Orthop Relat Res 193:16–19, 1985
8. Dennis S, Watkins R, Landaker S, Dillin W, Springer D: Com-
parison of disc space heights after anterior lumbar interbody
fusion. Spine 14:876–878, 1989
9. Ha KY, Shin JH, Kim KW, Na KH: The fate of anterior autog-
enous bone graft after anterior radical surgery with or without
posterior instrumentation in the treatment of pyogenic lumbar
spondylodiscitis. Spine 32:1856–1864, 2007
10. Hambly MF, Wiltse LL, Raghavan N, Schneiderman G,
Koenig C: The transition zone above a lumbosacral fusion.
Spine 23:1785–1792, 1998
11. Izumida S, Inoue S: [Assessment of treatment for low back
pain.] J Jpn Orthop Assoc 60:391–394, 1986 (Jpn)
12. Kanayama M, Hashimoto T, Shigenobu K, Harada M, Oha F,
Ohkoshi Y, et al: Adjacent-segment morbidity after Graf liga-
mentoplasty compared with posterolateral lumbar fusion. J
Neurosurg 95 (1 Suppl):5–10, 2001
13. Leong JC, Chun SY, Grange WJ, Fang D: Long-term results of
lumbar intervertebral disc prolapse. Spine 8:793–799, 1983
14. Min JH, Jang JS, Jung B, Lee HY, Choi WC, Shim CS, et al:
The clinical characteristics and risk factors for the adjacent
segment degeneration in instrumented lumbar fusion. J Spi-
nal Disord Tech 21:305–309, 2008
15. Miyakoshi N, Abe E, Shimada Y, Okuyama K, Suzuki T, Sato
K: Outcome of one-level posterior lumbar interbody fusion
for spondylolisthesis and postoperative intervertebral disc de-
generation adjacent to the fusion. Spine 25:1837–1842, 2000
16. Nagaosa Y, Kikuchi S, Hasue M, Sato S: Pathoanatomic
mechanisms of degenerative spondylolisthesis. A radiograph-
ic study. Spine 23:1447–1451, 1998
17. Nakai S, Yoshizawa H, Kobayashi S: Long-term follow-up study
of posterior lumbar interbody fusion. J Spinal Di sord 12:
293–299, 1999
18. Oda Y, Taguchi T, Fuchigami Y: [Facet joint: Etiological study
for lumbar degenerative spondylolisthesis.] Journal of Joint
Surgery 18:786–793, 1999 (Jpn)
19. Ohkohchi T, Ohwada T, Yamamoto T: [Long-term results of
Fig . 3. Schematic representation of L3–4 ASD caused by the dis-
traction of the L4–5 disc space (hypothesis). Distraction at L4–5 (gray
box) exerts compression on adjacent segments (arrows). The L3 –4
segment is the most susceptible to the L4–5 distraction.
J Neurosurg: Spine / Volume 12 / June 2010
A novel risk factor for adjacent-segment disease after PLIF
679
posterior lumbar interbody fusion for lumbar degenerative
disorders.] Rinsho Seikei Geka 35:519–526, 2000 (Jpn)
20. Ohkohchi T, Ohwada T, Yamamoto T: [The effect of PLIF for
lumbar degenerative spondylolisthesis on adjacent segments;
long term follow-up study.] Spine & Spinal Cord 11:657–
663, 1998 (Jpn)
21. Ohwada T, Ohkouchi T, Yamamoto T: [Long-term results of
PLIF with Steffee VSP system for degenerative spondylolis-
thesis.] Spine & Spinal Cord 17:193–200, 2004 (Jpn)
22. Okuda S, Iwasaki M, Miyauchi A, Aono H, Morita M, Yama-
moto T: Risk factors for adjacent segment degeneration after
PLIF. Spine 29:1535–1540, 2004
23. Park P, Garton HJ, Gala VC, Hoff JT, McGillicuddy JE: Ad-
jacent segment disease after lumbar or lumbosacral fusion:
review of the literature. Spine 29:1938–1944, 2004
24. Rohlmann A, Burra NK, Zander T, Bergmann G: Comparison
of the effects of bilateral posterior dynamic and rigid xa-
tion devices on the loads in the lumbar spine: a nite element
analysis. Eur Spine J 16:1223–1231, 2007
25. Satoh S, Kaneda K, Hosokawa Y, Fujitani M: [Decompres-
sion and stabilization for lumbar degenerative spondylolisthe-
sis with combined distraction and compression rod system:
long term follow-up results.] Clinical Orthopaedic Surgery
27:395–401, 1992 (Jpn)
26. Seitsalo S, Schlenzka D, Poussa M, Osterman K: Disc de-
generation in young patients with isthmic spondylolisthesis
treated operatively or conservatively: a long-term follow-up.
Eur Spine J 6:393–397, 1997
27. Sudo H, Oda I, Abumi K, Ito M, Kotani Y, Minami A: Bio-
mechanical study on the effect of ve different lumbar recon-
struction techniques on adjacent-level intradiscal pressure and
lamina strain. J Neurosurg Spine 5:150–155, 2006
28. Trouillier H, Birkenmaier C, Rauch A, Weiler C, Kauschke
T, Reor HJ: Posterior lumbar interbody fusion (PLIF) with
cages and local bone graft in the treatment of spinal stenosis.
Acta Orthop Belg 72:460– 466, 2006
29. Vanharanta H, Floyd T, Ohnmeiss DD, Hochschuler SH, Guy-
er RD: The relationship of facet tropism to degenerative disc
disease. Spine 18:1000–1005, 1993
30. Weishaupt D, Zanetti M, Boos N, Hodler J: MR imaging and
CT in osteoarthritis of the lumbar facet joints. Skeletal Ra-
diol 28:215–219, 1999
31. Yamamoto T, Kadowaki T, Ohta N, Ohwada T: [PLIF for
degenerative spondylolisthesis–effects of additional pedicle
screw xation systems on preventing collapse of graft.] Clini-
cal Orthopaedic Surgery 25:487–494, 1990 (Jpn)
32. Yamamoto T, Kadowaki T, Ohwada T: [Posterior lumbar in-
terbody fusion-Cloward method.] Spine & Spi nal Cord 5:
355–362, 1992 (Jpn)
33. Yamamoto T, Ohkohchi T, Ohwada T: Clinical and radiologi-
cal results of PLIF for degenerative spondylolisthesis. J Mus
Res 2:181–195, 1998
34. Yasukawa Y: [Age changes in the lumbar spine; radiological
follow-up studies over more than ten years.] J Jpn Orthop
Assoc 68:854–863, 1994 (Jpn)
Manuscript submitted November 11, 2008.
Accepted December 4, 2009.
Address correspondence to: Takashi Kaito, M.D., Department
of Orthopedics, National Hospital Organization Osaka Minami
Medical Center, 2-1 Kidohigashi, Kawachinagano, Osaka 586-8521,
Japan. email: t-kaito@leto.eonet.ne.jp.