Arch Orthop Trauma Surg
DOI 10.1007/s00402-017-2757-0
ORTHOPAEDIC SURGERY
Changes of spinopelvic parameters in different positions
Soo‑An Park1 · Dai‑Soon Kwak2 · Ho‑Jung Cho2 · Dong‑Uk Min1
Received: 27 April 2017
© Springer-Verlag GmbH Germany 2017
Abstract
Introduction To observe changes of spinopelvic param-
eters and the presence of pelvic incidence (PI) variation in
different positions, and the accuracy of PI compared with
CT scan.
Materials and methods Patients with standing whole-
spine radiograph, CT scan of the pelvic bone, and MRI of
the lumbar spine done within a few days were included.
The pelvic [pelvic tilt (PT), sacral slope (SS), and PI] and
spinal [lumbar lordosis (LL)] parameters were measured by
two different observers.
Results The PIs from radiograph were significantly greater
than those from CT in both observers. By adopting the
upper limit of the confidence interval and the agreement of
two observers on grouping, patients were categorized into
two subgroups (SG1, with less PI change; SG2, with higher
PI change). The PT and LL values decreased, whereas SS
increased significantly from standing to supine positions in
SG1. Significantly decreased PT and PI from standing to
supine were observed in SG2. All pelvic parameters and the
sagittal vertical axis on radiograph, and the LL amount on
MRI were significantly greater in SG2 than in SG1.
Conclusions Majority of patients demonstrated alignment
changes of unchanged PI with decreased PT and LL, and
* Dai‑Soon Kwak
daisoon@catholic.ac.kr
1
Department of Orthopedic Surgery, Uijeongbu St.
Mary’s Hospital, College of Medicine, The Catholic
University of Korea, 271, Cheonbo‑Ro, Uijeongbu‑si,
Gyeonggi‑do 11765, Republic of Korea
2
Department of Anatomy, Catholic Institute for Applied
Anatomy, College of Medicine, The Catholic University
of Korea, 222, Banpo‑daero, Seocho‑gu, Seoul 06591,
Republic of Korea
increased SS from standing to supine; however, decreased
PT and PI and fixed SS and LL were also demonstrated.
Patients with higher PI change have high values in three
pelvic parameters and sagittal vertical axis, and fixed LL.
Keywords Spinopelvic parameter · Standing · Supine
position · CT scan of pelvic bone
Introduction
For the spinopelvic parameters, we can describe some limi-
tations. First, these parameters are measured on standing
whole-spine radiograph [1], but standing does not represent
the postures used in daily activities, and it does not show
alignment and deformity of spine and pelvis in non-weight-
loading condition.
Second, the pelvic incidence (PI) is known as a constant
value [2], but can change after a disruption in the sacro-
iliac joint (SIJ), such as in cases of spinopelvic dissociation
injury or long fusion in patients with SIJ laxity [3, 4]. Thus,
when evaluating a big valued PI, it is difficult to differenti-
ate whether it is from stable but anatomically wide SIJ, SIJ
laxity, or an error in measurement.
Third, the radiographic parameters are affected by the
direction and distance between patient and radiographic
system. Therefore, it is hard to estimate the error range of
these parameters comparing to the true values. Compared
with the X-ray, the CT scan has a benefit in quantitative
measurement, because CT image is not affected by the
radiographic factors [5, 6]. In addition, when enhancing
the transparency of 3D CT image, we can collect the image
like the plain radiograph with less image-dependent errors.
This study was to determine how the spinopelvic align-
ment changes and whether the PI changes from standing with
13
Vol.:(0123456789)
 Arch Orthop Trauma Surg

weight-loading to supine with non-weight-loading conditions,
and to test the accuracy of PI compared with the value obtained
from CT scan, under the hypothesis that PI is not different
between different imaging tests, positions, and observers.
testing alignment change of spine and pelvis from stand-
ing to supine and for testing the presence of PI variation
for different weight-loading conditions. Measurements per-
formed by two different observers were to test the observer-
dependent reliability.

Materials and methods Measurements on CT scan of pelvic bone

Patients
One hundred and four patients (58.5 ± 17.0 years; 40
men and 64 women) with lumbar degenerative disease
who underwent standing whole-spine radiography, CT
scan of the pelvic bone, and MRI of the lumbar spine in
supine position done within a few days of each other with-
out undergoing any invasive procedures between imaging
tests were selected. All patients demonstrated symptoms
of lumbar spinal stenosis and corresponding radiographic
findings. Patients of lumbar spinal stenosis by other than
degenerative causes and those with age less than 20 years
were excluded in this study.
Experimental setup
The experiment compared the spinopelvic parameters
between standing radiograph and supine CT/MRI for
The CT image of the pelvic cage with two pelvic bones
and one sacrum was collected as DICOM files from each
patient. The image files were reconstructed into 3-D model
of the pelvic cage for each patient using Mimics software
(version 18; Materialise, Leuven, Belgium). After changing
the opacity of the displayed bone to maximally transparent,
the 3-D model was placed in the sagittal plane to demon-
strate the true lateral view to simulate the plain radiograph.
The pelvic cage was bisected at the mid-sagittal plane and
the opposite portion was hidden from the screen to clearly
demonstrate each acetabulum and the superior endplate
of S1. The acetabulum was marked on the pelvic bone of
the ipsilateral side by a maximal circle, and the endplate
of S1 was marked on the mid-sagittal plane of the sacrum
(Fig. 1).
Each 2-D lateral image of the pelvic cage with anatomi-
cal landmarks for both sides was imported into the computer-
aided design (CAD) software Draftsight 2015 (Simulia,

Fig. 1  Techniques of image

processing and measurement of
spinopelvic parameters. a Mark-
ing the midline on sacrum and
symphysis pubis bisecting the
pelvic cage. b After adjusting
the transparency, a 3-D model
was placed in the sagittal plane
with both sides overlapped
as much as possible. c Each
acetabulum and the superior
endplate of S1 were marked
on the 2-D image of the pelvic
cage with a maximal circle and
a straight line after hiding the
contralateral side, respectively.
d Pelvic parameters were
calculated with the collected
positional data for acetabular
circles and S1 endplate in the
sagittal plane

13
Arch Orthop Trauma Surg

France), and the pelvic tilt (PT), sacral slope (SS), and PI
were calculated using the positional data of the anatomical
landmarks (Fig. 1).
Measurements on whole‑spine radiograph and lumbar
spine MRI
On standing whole-spine radiograph, PT, SS, and PI were
measured for the pelvic parameters, and lumbar lordosis
(LL, angle between superior endplates of L1 and S1), tho-
racic kyphosis (TK, angle between the superior endplate
of T4 and the inferior endplate of T12), sagittal vertical
axis (SVA), and PI–LL were measured/calculated for the
spinal parameters. The LL value on MRI was measured
using the mid-sagittal view in the same manner used in
plain radiograph. The PI–LL value in the supine position
was calculated using PI from CT scan and LL from MRI
(Figs. 2, 3).

Fig. 2  Measurements of
spinopelvic parameters on
standing whole-spine radio-
graph, mid-sagittal image of the
lumbar spine MRI in supine,
and transparent lateral image
of pelvic bone CT in supine. A
61-year-old male patient in sub-
group 1 demonstrated decreased
PT, increased SS, and decreased
LL values from standing to
supine positions. Therefore, PI
was almost constant in different
positions

13
 Arch Orthop Trauma Surg

Fig. 3  90-year-old female

patient in subgroup 2 dem-
onstrated unchanged SS and
LL and decreased PT. There-
fore, the PI and PI–LL values
decreased from standing to
supine. This patient demon-
strated widely spread SIJ, some
stooping, and thoracolumbar
insufficiency fracture

Statistical analyses
To analyze interobserver reliability, a pairwise t test was
performed for every parameter included (Table 1). To test
the difference in PI between radiograph and CT scan, a pair-
wise t test was performed in each data set of observers A and
B. Using these results, patients were divided into group 1
[pairwise difference ≤the upper limit of the 95% confidence
interval (CI)] and group 2 (>the upper limit of the CI) in
each observer. If both observers agreed on the grouping, the
patients were regrouped as subgroups 1 (SG1) and 2 (SG2)
(Table 2).
To compare the clinical characteristics of the patients
between subgroups, one-way ANOVA and χ2 tests
were performed in accordance with the type of variable
(Table 3).
To compare the parametric differences between the two
subgroups, MANOVA were performed for pelvic (PT, SS,

13
Arch Orthop Trauma Surg

Table 1  Pairwise difference Parameters N Observer PW difference Correlation

and correlation of each
parameter between observers A B Mean (SD) 95% CI
A and B
Mean (SD) Mean (SD) Lower Upper ICC P

Pelvic (radiograph)
PT (°) 71 18.6 (9.9) 18.5 (9.6) 0.1 (1.9) −0.3 0.6 0.981 <0.001*
SS (°) 71 34.7 (9.8) 35.7 (10.1) −0.9 (2.8) −1.6 −0.2 0.960 <0.001*
PI (°) 71 53.3 (11.0) 54.1 (10.8) −0.8 (4.0) −1.7 0.2 0.933 <0.001*
Pelvic (CT)
PT (°) 71 12.0 (6.8) 11.9 (6.3) 0.1 (2.6) −0.5 0.8 0.922 <0.001*
SS (°) 71 39.6 (8.3) 38.2 (8.0) 1.4 (3.5) 0.5 2.2 0.907 <0.001*
PI (°) 71 51.6 (9.5) 50.1 (9.7) 1.5 (2.3) 1.0 2.1 0.971 <0.001*
Spinal (radiograph)
LL (°) 71 −42.0 (15.5) −43.1 (15.3) 1.1 (3.5) 0.3 1.9 0.974 <0.001*
TK (°) 71 32.3 (12.7) 33.2 (12.2) −0.8 (3.8) −1.7 0.1 0.954 <0.001*
PI–LL (°) 71 11.3 (15.4) 11.0 (14.8) 0.3 (2.9) −0.4 1.0 0.983 <0.001*
SVA (mm) 45 38.2 (34.0) 37.5 (33.9) 0.7 (4.1) −0.5 1.9 0.993 <0.001*
Spinal (CT/MRI)
LL (°) 71 −38.7 (11.2) −38.3 (11.2) −0.4 (4.0) −1.3 0.6 0.936 <0.001*
PI–LL (°) 71 12.9 (10.1) 11.8 (10.2) 1.2 (4.3) 0.2 2.2 0.912 <0.001*

* The correlation of each parameter between observers A and B was significant

Table 2  Pairwise differences Param- N Radiograph CT PW difference Correlation

and correlations of PI eter/
between standing whole-spine observer Mean (SD) 95% CI P ICC P
radiograph and supine CT scan
of pelvic bone in the data sets of Mean (SD) Mean (SD) Lower Upper
observers A and B
PI (°)
By A 104 52.8 (11.0) 50.6 (10.1) 2.3 (5.0) 1.3 3.2 <0.001* 0.893 <0.001†
By B 104 53.2 (10.9) 49.1 (10.3) 4.0 (3.9) 3.3 4.8 <0.001* 0.934 <0.001†

* The difference of PI between radiograph and CT scan was significant

†
The correlation of PI between radiograph and CT scan was significant

and PI) and spinal (LL, TK, SVA, and PI–LL) parameters
from standing radiograph, the same pelvic parameters from
CT scan, and the spinal parameters (LL, PI–LL) from MRI/
CT scan as dependent variables and subgroup as independ-
ent variable (Table 4). To analyze the changes in spinopel-
vic alignment from standing to supine positions, pairwise
t tests were performed for each parameter in SG1 and SG2
(Table 5).
Results
Reliability of spinopelvic parameters
All parameters from observers A and B were significantly
correlated, and the intraclass correlation coefficients (ICC)
for all were >0.9. The pairwise mean differences between
two observers for all angular parameters on radiograph and
MRI/CT scan were lower than 2°. The mean difference for
SVA was 0.7 mm (Table 1).
The PI from the radiograph to CT scan is highly corre-
lated (ICC: observer A, 0.893; observer B, 0.934), but the
ICC values are mostly lower than those from variabilities
between two observers. The PI values were significantly
higher on the radiograph (observer A: mean 52.8° vs.
50.6°; mean difference 2.3°; 95% CI 1.3°–3.2°; P < 0.001;
observer B: 53.2° vs. 49.1°; 4.0°; 3.3°–4.8°; P < 0.001)
compared with those on the CT scan in both observers
(Table 2).
Patients in subgroups 1 and 2
Fifty-three patients were included in SG1
(60.6 ± 12.6 years; 20 men and 33 women) and 18 in SG2

13
 Arch Orthop Trauma Surg

(69.4 ± 13.6 years; 2 men and 16 women). Patients in SG2 Table 3  Clinical comparison of patients for the two subgroups
were significantly older, included more women, and dem- SG1 SG2 P
onstrated more thoracolumbar insufficiency fractures than
Mean SD Mean SD
those in SG1. Patients in SG1 and SG2 are not significantly
different in levels of spinal stenosis and presence of spinal Age (years) 60.6 12.6 69.4 13.6 0.014*
deformity, spondylolisthesis, and previous lumbar spinal Levels of stenosis (n) 2.7 1.0 2.9 1.2 0.510
surgery (Table 3). N % N %

Comparison of spinopelvic parameter for the two Sex

subgroups Females 33 62 16 89 0.030†
Males 20 38 2 11
All pelvic parameters (PT: 23.2° vs. 17.0° and 22.7° vs. Spinal stenosis
17.0°; SS: 39.2° vs. 33.2° and 39.8° vs. 34.2°; PI: 62.6° No 0 0 0 0 NA
vs. 50.2° and 62.5° vs. 51.3°) and SVA (60.1 vs. 32.9 and Yes 53 100 18 100
71.6 vs. 31.9 mm) on standing radiograph were signifi- Spinal deformity (≥30° Cobb angles)
cantly greater in SG2 than in SG1 in both observers. The No 48 91 14 78 0.158
LL values on supine MRI (−36.8° vs. −44.3° and −36.4° Yes 5 9 4 22
vs. −43.8°) were significantly greater in SG1 than in SG2 Spondylolisthesis
in both observers (Table 4). No 44 83 16 89 0.432
Yes 9 17 2 11
Changes in spinopelvic parameters from standing Thoracolumbar insufficiency fracture
to supine positions in SG1 No 49 93 11 61 0.004†
Yes 4 7 7 39
PT decreased (observer A: mean 17.0° vs. 11.6°; mean Previous lumbar spinal surgery
change 5.4°; 95% CI 3.2°–7.6°; observer B: 17.0° vs. No 50 94 14 78 0.064
11.3°; 5.7°; 4.1°–7.4°) and SS increased (observer A: Yes 3 6 4 22
32.2° vs. 39.3°; −6.1°; −8.4° to −3.8°; observer B: 34.2° * The difference of parametric variable between the two subgroups
vs. 37.8°; −3.6°; −5.5° to −1.7°) significantly from stand- was significant
ing to supine positions in both observers. In observer A, PI †
The difference of non-parametric variable between the two sub-
was not significantly changed. However, in observer B, the groups was significant
value of PI was significantly decreased from radiograph to
CT scan, but the difference was low (mean difference 2°;
95% CI 1.4°–2.9°). For the spinal parameters, PI significantly decreased
For the spinal parameters, PI was unchanged (less but LL, and therefore, PI–LL decreased significantly in
changed) and the amount of LL significantly decreased both observers (observer A: mean 15.5° vs. 9.4°; mean
(observer A: mean −40.3° vs. −36.8°; mean change change 6.1°; 95% CI 0.7°–11.6°; observer B: 15.2° vs.
−3.5°; 95% CI −6.1° to −1.0°; observer B: −41.7° vs. 9.0°; 6.2°; 0.8°–11.6°) (Tables 4, 5; Figs. 3, 4).
−36.4°; −5.2°; −7.4° to −3.0°), and therefore, the value
of PI–LL increased significantly (observer A: 9.9° vs.
14.2°; −4.2°; −6.9° to −1.6°; observer B: 9.6° vs. 12.7°; Discussion
−3.1°; −5.2° to −1.0°) from standing to supine positions
in both observers (Tables 4, 5; Figs. 2, 4). All spinopelvic parameters demonstrated pairwise differ-
ence between two observers with <2° and 2 mm and ICC
Changes in spinopelvic parameters from standing >0.9. Thus, we can conclude that the spinopelvic param-
to supine positions in SG2 eters included in this study are highly reliable in interob-
server errors.
Both observers noted a significant decrease in PI (observer When considering intertest variability of PI between
A: mean 62.6° vs. 53.6°; mean change 9.0°; 95% CI radiograph and CT scan, the PIs for two imaging studies
7.0°–10.9°; observer B: 62.5° vs. 52.8°; 9.7°; 8.1°–11.3°), were highly correlated, with ICC of 0.893 and 0.934, and
because the SS did not change for the significant decrease were measured consistently greater on radiograph than on
in PT (observer A: 23.4° vs. 13.3°; 10.0°; 6.3°–13.8°; CT scan in both observers (95% CI of mean PI difference
observer B: 22.7° vs. 13.6°; 9.1°; 5.6°–12.6°). between radiograph and CT scan, observer A: 1.3°–3.2°;

13
Arch Orthop Trauma Surg

Table 4  Comparison of each Parameters SG1 SG2 P (SG1 vs. SG2)

parameter for two subgroups A/B
using values from observers A N Mean (SD) N Mean (SD)
and B
Observers A/B Observers A/B

Pelvic (radiograph)
PT (°) 53 17.0 (9.0)/17.0 (8.9) 18 23.4 (10.9)/22.7 (10.6) 0.017*/0.030†
SS (°) 53 33.2 (9.2)/34.2 (9.7) 18 39.2 (10.4)/39.8 (10.2) 0.023*/0.040†
PI (°) 53 50.2 (8.4)/51.3 (8.7) 18 62.6 (12.7)/62.5 (12.1) <0.001*/<0.001†
Pelvic (CT)
PT (°) 53 11.6 (7.1)/11.3 (6.4) 18 13.3 (6.0)/13.6 (5.9) 0.352/0.189
SS (°) 53 39.3 (8.2)/37.8 (7.5) 18 40.3 (8.9)/39.3 (9.6) 0.664/0.514
PI (°) 53 50.9 (8.8)/49.1 (8.9) 18 53.6 (11.4)/52.8 (11.8) 0.296/0.166
Spinal (radiograph)
LL (°) 53 −40.3 (14.8)/−41.7 (14.8) 18 −47.1 (16.7)/−47.3 (16.6) 0.108/0.176
TK (°) 53 30.7 (12.9)/33.1 (13.1) 18 37.3 (10.9)/33.5 (10.7) 0.055/0.926
PI–LL (°) 53 9.9 (15.1)/9.6 (14.1) 18 15.5 (15.9)/15.2 (16.3) 0.183/0.171
SVA (mm) 36 32.9 (30.3)/31.9 (29.6) 9 60.1 (39.8)/71.6 (55.1) 0.029*/0.004†
Spinal (CT/MRI)
LL (°) 53 −36.8 (10.5)/−36.4 (10.5) 18 −44.3 (11.6)/−43.8 (11.5) 0.013*/0.014†
PI–LL (°) 53 14.2 (10.1)/12.7 (9.9) 18 9.4 (9.2)/9.0 (10.8) 0.082/0.182

* The difference of each parameter between SG1 and SG2 in the data set of observer A was significant
†
The difference of each parameter between SG1 and SG2 in the data set of observer B was significant

Table 5  Pairwise difference of Subgroups Parameters N PW difference

each parameter from radiograph
to MRI/CT scan in subgroups 1 Mean (SD) 95% CI (lower to upper) P
and 2 using values measured by
observers A and B Observers A/B Observers A/B A/B

SG1 PT (°) 53 5.4 (8.1)/5.7 (6.0) 3.2 to 7.6/4.1 to 7.4 <0.001*/<0.001†

SS (°) 53 −6.1 (8.3)/−3.6 (6.9) −8.4 to −3.8/−5.5 to −1.7 <0.001*/0.001†
PI (°) 53 −0.7 (3.8)/2.1 (2.8) −1.7 to 0.3/1.4 to 2.9 0.173/<0.001†
LL (°) 53 −3.5 (9.3)/−5.2 (8.0) −6.1 to −1.0/−7.4 to −3.0 0.008*/<0.001†
PI–LL (°) 53 −4.2 (9.6)/−3.1 (7.7) −6.9 to −1.6/−5.2 to −1.0 0.002*/0.005†
SG2 PT (°) 18 10.0 (7.5)/9.1 (7.1) 6.3 to 13.8/5.6 to 12.6 <0.001*/<0.001†
SS (°) 18 −1.1 (7.1)/0.6 (6.5) −4.6 to 2.5/−2.7 to 3.8 0.527/0.720
PI (°) 18 9.0 (4.0)/9.7 (3.2) 7.0 to 10.9/8.1 to 11.3 <0.001*/<0.001†
LL (°) 18 −2.8 (10.7)/−3.5 (10.0) −8.1 to 2.5/−8.5 to 1.5 0.279/0.157
PI–LL (°) 18 6.1 (11.0)/6.2 (10.8) 0.7 to 11.6/0.8 to 11.6 0.030*/0.027†

* The pairwise difference of each parameter between standing radiograph and supine MRI/CT scan was
significant in the data set of observer A
†
The pairwise difference of each parameter between standing radiograph and supine MRI/CT scan was
significant in the data set of observer B

observer B: 3.3°–4.8°). Based on the upper limit of CI in
observer B, PI appears to be measured by up to 5° greater
on the conventional radiograph than on CT scan.
However, when differentiating subgroups and looking
at the intertest difference of PI in SG2, PI was about 10°
greater on radiograph than on CT scan, with narrow CIs in
both observers, which cannot be explained by variabilities
from tests, observers, and/or measurement tools.
In SG2, the SS did not change with the decrease in PT
from standing to supine, and therefore, the values of PI
decreased (observer A: from 62.6°–53.6°; observer B: from
62.5°–52.8°) in both observers. For the spinal parameters,

13

Fig. 4  Pairwise changes in PT,
SS, PI, LL, and PI–LL from
standing radiograph (red dot) to
supine MRI/CT scan (blue dot)
in subgroups 1 and 2 in the data
sets of observer A
decreased PI and unchanged LL led to a decrease in PI–LL
from standing to supine in SG2.
Patients in SG1 demonstrated a concomitant PT decrease
with the SS increase in both observers, and therefore, PI did
not change from standing to supine positions (in observer
A). Although PI decreased significantly in observer B, the
amount of change was quite small (mean difference 2°;
95% CI 1.4°–2.9°). The different PI changes in two observ-
ers in SG1 were due to the greater variability of SS between
observers in estimating the vertebral endplate than that of
13
Arch Orthop Trauma Surg
PT on both radiograph and CT scan [7, 8]. For the spinal
parameters in SG1, PI was unchanged and the amount of
LL decreased from standing to supine, and therefore, the
value of PI–LL increased in both observers.
Patients in SG2 have greater PI, PT, SS, and SVA during
standing and a greater amount of LL during supine posi-
tion when compared with those in SG1 in both observers.
High values in all pelvic parameters (PI, PT and SS) indi-
cate pelvic retroversion with more vertically rotated sacral
endplate, resulting in horizontally spread SIJ. Patients with
Arch Orthop Trauma Surg
widely spread SIJ may be vulnerable to axial load in stand-
ing. Higher LL during supine position in SG2 than in SG1
might be from unchanged LL from standing to supine,
and that is observed in patients with degenerative spinal
deformity with less mobile or fixed lumbar segments [9].
Clinically, patients in SG2 were older, mostly women, and
had more thoracolumbar insufficiency fractures than those
in SG1.
The previous studies using positional MRI have reported
dynamic changes in LL from standing to supine in healthy
individuals [10, 11] and in LBP patients [12, 13]. The
amount of LL increased with the increase in lumbar exten-
sion in standing, and the changes in LL between positions
were not different in both groups; the only difference was
the lower amount of LL (both on standing and supine
MRIs) in LBP patients compared with healthy individuals.
Decreased LL in the supine position from standing in the
current SG1 matches with the findings of previous stud-
ies [12, 13] and are due to mobile lumbar segments. An
unchanged LL for different positions in SG2 is unusual,
even with the presence of lumbar degenerative disease, and
this may be because patients in SG2 have the character-
istics of lumbar degenerative deformity with less or fixed
segmental motion [9].
The different values of the same parameters observed
in two different imaging studies were from two differ-
ent imaging modalities and from two different positions
(standing and supine). PT, SS, and LL are affected by
weight-loading, postures of lower extremity joints, and
forward bending of the spinal column [14, 15], but the
constant value of PI should not be affected by any factors
[2]; hence, the differences in PT, SS, and LL between radi-
ograph and MRI/CT scan in this study were interpreted as
changes from standing to supine positions, but the differ-
ence in PI was interpreted as an unusual or non-physio-
logic finding.
The reason for high PI decrease from standing to supine
that was observed in the current SG2 could be explained
by the fact that the PI in SG2 on plain radiograph might
be overestimated, because the pelvic bone was more retro-
verted (higher PT) and the sacral endplate was more verti-
cally oriented (higher SS) in standing compared with those
in SG1. However, if PI change between standing radiograph
and supine CT images exceeds the expected range consist-
ently in multiple observers, this may indicate the possibility
that weight-loading in standing increased the PI by affect-
ing the unstable SIJ.
In conclusion, the spinopelvic parameters are highly
reliable in interobserver errors, but might be overes-
timated by up to 5° in radiograph compared with CT
scan especially in PI. Alignment changes in the form of
decreased PT, increased SS, and decreased LL should be
the normal physiologic event from standing to supine;
however, decreased PT and PI and fixed SS and LL were
also demonstrated. Patients with higher PI change have
high values in all three pelvic parameters, high SVA, and
fixed LL, and they are usually older and women and more
likely to have thoracolumbar insufficiency fractures com-
pared with their counterpart.
The PI with less image-dependent errors and with no
effects of weight-loading and SIJ could be collected by
analyzing 2D reformatted CT image of the pelvic bone.
Because the PI is a reference value for LL correction
[16], comparing the values of PI from the radiograph and
CT images may provide the information about the SIJ
pathology preoperatively and should be helpful in proper
planning of LL reconstruction. However, further clinical
correlation should be evaluated in future studies.
Compliance with ethical standards
Conflict of interest The authors declare that there is no conflict of
interest regarding the publication of this paper.
Funding There is no funding source.
Ethical approval All procedures performed in studies involving
human participants were in accordance with the ethical standards of
the institutional and/or national research committee and with the 1964
Helsinki declaration and its later amendments or comparable ethical
standards. This study was approved by the Institutional Review Board
(UC16RIMI0023).
References
1. Marks M, Stanford C, Newton P (2009) Which lateral
radiographic positioning technique provides the most reli-
able and functional representation of a patient’s sagittal bal-
ance? Spine (Phila Pa 1976) 34(9):949–954. doi:10.1097/
BRS.0b013e318199650a
2. Mac-Thiong JM, Roussouly P, Berthonnaud E, Guigui P (2011)
Age- and sex-related variations in sagittal sacropelvic morphol-
ogy and balance in asymptomatic adults. Eur Spine J 20(Suppl
5):572–577. doi:10.1007/s00586-011-1923-2
3. Hart RA, Badra MI, Madala A, Yoo JU (2007) Use of pelvic
incidence as a guide to reduction of H-type spino-pelvic disso-
ciation injuries. J Orthop Trauma 21(6):369–374. doi:10.1097/
BOT.0b013e31806dd959
4. Charles YP, Yu B, Steib JP (2016) Sacroiliac joint luxation
after pedicle subtraction osteotomy: report of two cases and
analysis of failure mechanism. Eur Spine J 25(Suppl 1):63–74.
doi:10.1007/s00586-015-4094-8
5. Hauser CJ, Visvikis G, Hinrichs C, Eber CD, Cho K, Lavery
RF, Livingston DH (2003) Prospective validation of computed
tomographic screening of the thoracolumbar spine in trauma. J
Trauma 55(2):228–234. doi:10.1097/01.TA.0000076622.19246.
CF (discussion 234–225)
6. Brandt MM, Wahl WL, Yeom K, Kazerooni E, Wang SC (2004)
Computed tomographic scanning reduces cost and time of com-
plete spine evaluation. J Trauma 56(5):1022–1026 (discussion
1026–1028)
13

7. Dimar JR 2nd, Carreon LY, Labelle H, Djurasovic M, Weiden-
baum M, Brown C, Roussouly P (2008) Intra- and inter-observer
reliability of determining radiographic sagittal parameters of
the spine and pelvis using a manual and a computer-assisted
methods. Eur Spine J 17(10):1373–1379. doi:10.1007/
s00586-008-0755-1
8. Frobin W, Brinckmann P, Biggemann M, Tillotson M, Burton
K (1997) Precision measurement of disc height, vertebral height
and sagittal plane displacement from lateral radiographic views
of the lumbar spine. Clin Biomech (Bristol, Avon) 12(Suppl
1):S1–S63
9. Suk S-I, Chung E-R, Lee S-M, Lee J-H, Kim S-S, Kim J-H
(2005) Posterior vertebral column resection in fixed lumbosacral
deformity. Spine (Philadelphia, Pa 1976) 30(23):E703–E710
10. Mauch F, Jung C, Huth J, Bauer G (2010) Changes in the lum-
bar spine of athletes from supine to the true-standing position in
magnetic resonance imaging. Spine (Phila Pa 1976) 35(9):1002–
1007. doi:10.1097/BRS.0b013e3181bdb2d3
11. Hirasawa Y, Bashir WA, Smith FW, Magnusson ML, Pope MH,
Takahashi K (2007) Postural changes of the dural sac in the lum-
bar spines of asymptomatic individuals using positional stand-up
magnetic resonance imaging. Spine (Phila Pa 1976) 32(4):E136–
E140. doi:10.1097/01.brs.0000255202.94153.ca
12. Tarantino U, Fanucci E, Iundusi R, Celi M, Altobelli S, Gasbarra
E, Simonetti G, Manenti G (2013) Lumbar spine MRI in upright
13
Arch Orthop Trauma Surg
position for diagnosing acute and chronic low back pain: sta-
tistical analysis of morphological changes. J Orthop Traumatol
14(1):15–22. doi:10.1007/s10195-012-0213-z
13. Hansen B, Bendix T, Grindsted J, Bliddal H, Christensen R,
Hansen P, Riis RGC, Boesen M (2015) Effect of lumbar disc
degeneration and low-back pain on the lumbar lordosis in supine
and standing: a cross-sectional MRI study. Spine (Philadelphia,
Pa 1976) 40(21):1690–1696
14. Protopsaltis T, Schwab F, Bronsard N, Smith J, Klineberg
E, Mundis G, Ryan D, Hostin R, Hart R, Burton D, Ames C,
Shaffrey C, Bess S, Errico T, Lafage V (2014) The T1 pelvic
angle, a novel radiographic measure of global sagittal deform-
ity, accounts for both spinal inclination and pelvic tilt and cor-
relates with health-related quality of life. J Bone Jt Surg Am
96(19):1631–1640
15. Neuschwander TB, Cutrone J, Macias BR, Cutrone S, Murthy
G, Chambers H, Hargens AR (2010) The effect of backpacks
on the lumbar spine in children: a standing magnetic resonance
imaging study. Spine (Phila Pa 1976) 35(1):83–88. doi:10.1097/
BRS.0b013e3181b21a5d
16. Ames CP, Smith JS, Scheer JK, Bess S, Bederman SS, Deviren
V, Lafage V, Schwab F, Shaffrey CI (2012) Impact of spinopelvic
alignment on decision making in deformity surgery in adults: a
review. J Neurosurg Spine 16(6):547–564. doi:10.3171/2012.2.S
PINE11320