Eur Spine J (2000) 9 : 242–248
`© Springer-Verlag 2000
`
`ORIGINAL ARTICLE
`
`S. H. Zhou
`I. D. McCarthy
`A. H. McGregor
`R. R. H. Coombs
`S. P. F. Hughes
`
`Geometrical dimensions
`of the lower lumbar vertebrae –
`analysis of data from digitised CT images
`
`Received: 9 November 1998
`Revised: 14 January 2000
`Accepted: 26 January 2000
`
`S. H. Zhou (Y) · I. D. McCarthy ·
`A. H. McGregor · R. R. H. Coombs ·
`S. P. F. Hughes
`Department of Orthopaedic
`and Trauma Surgery,
`Imperial College School of Medicine,
`Charing Cross Hospital,
`Fulham Palace Road,
`London W6 8RF, UK
`e-mail: s.zhou@ic.ac.uk,
`Tel.: +44-208-8461678,
`Fax: +44-208-8461439
`
`Introduction
`
`Abstract The precise dimensions of
`the lumbar vertebrae and discs are
`critical for the production of appro-
`priate spinal implants. Unfortunately,
`existing databases of vertebral and
`intervertebral dimensions are limited
`either in accuracy, study population
`or parameters recorded. The objec-
`tive of this study is to provide a large
`and accurate database of lumbar
`spinal characteristics from 126 digi-
`tised computed tomographic (CT)
`images, reviewed using the Picture
`Archiving Communication System
`(PACS) coupled with its internal
`measuring instrumentation. These
`CT images were obtained from pa-
`tients with low back pain attending
`the spinal clinic at the Hammersmith
`Hospitals NHS Trust. Measurements
`of various aspects of vertebral di-
`mensions and geometry were re-
`corded, including vertebral and inter-
`vertebral disc height. The results
`from this study indicated that the
`depth and width of the vertebral end-
`plate increased from the third to the
`fifth lumbar vertebra. Anterior verte-
`bral height remained the same from
`
`the third to the fifth vertebra, but the
`posterior vertebral height decreased.
`Mean disc height in the lower lum-
`bar segments was 11.6 ± 1.8 mm for
`the L3/4 disc, 11.3 ± 2.1 mm for the
`L4/5, and 10.7 ± 2.1 mm for the
`L5/S1 level. The average circumfer-
`ence of the lower endplate of the
`fourth lumbar vertebra was 141 mm
`and the average surface area was
`1492 mm2. An increasing pedicle
`width from a mean of 9.6 ± 2.2 mm
`at L3 through to 16.2 ± 2.8 mm at
`L5 was noted. A comprehensive
`database of vertebral and interverte-
`bral dimensions was generated from
`378 lumbar vertebrae from 126 pa-
`tients measured with a precise digital
`technique. These results are invalu-
`able in establishing an anthropomet-
`ric model of the human lumbar
`spine, and provide useful data for
`anatomical research. In addition this
`is important information for the sci-
`entific planning of spinal surgery and
`for the design of spinal implants.
`
`Key words Lumbar vertebrae ·
`Anatomical dimensions · Spine
`
`Accurate and comprehensive anthropometric data for the
`lumbar spine vertebrae, a frequent site for implantation
`surgery, are incomplete at present. Information on the pre-
`cise dimensions of the lower lumbar vertebrae is, how-
`ever, essential, for the rational design and development of
`
`spinal implants and instrumentation such as pedicle
`screws and, in particular, with the evolution towards ro-
`botic surgery. Previous studies have depended on direct
`measurements from plain X-ray films [9, 12, 13, 23], or
`from computed tomographic (CT) scans [8, 11, 26, 34,
`36]. A few reports have involved the analysis of cadaveric
`specimens [1, 7, 24, 27, 29]. The value of the data has de-
`pended on the number of samples and the accuracy of
`
`MSD 1012
`
`1
`
`

`

`243
`
`lower endplates for each vertebra and were studied from the third
`lumbar vertebra to the upper sacrum. A study using slices of 1 mm
`or less would have provided more precise data on cortical thick-
`ness, but the increased radiation dose could not be justified on clin-
`ical grounds in a study of living patients. In addition, a lateral to-
`mogram of the whole spine was obtained. The pixel of the CT scan
`was no greater than 0.11 mm in size, and the zoom factor was 4.5.
`The images were digitised and stored on the Picture Archiving
`Communication System (PACS). This is a computerised system
`for recording and storing radiographic images, permitting storage
`of large numbers of images, and allowing access from any net-
`worked station. In addition to these storage facilities, it also incor-
`porates image enhancement and manipulation tools such as magni-
`fication and rotation. The software of the PACS system also incor-
`porates a sensitive measuring tool. To measure the distance be-
`tween two points, a cursor is positioned using the mouse over an
`initial reference point. The cursor is then moved to the second ref-
`erence point by dragging the mouse. When the mouse button is re-
`leased, the distance between the two points is automatically dis-
`played in the information box, reflecting not only a measurement
`from the CT film [20] but also the actual size of the vertebral body
`in the plane of the slice. Nine parameters were measured from the
`cross-sectional images and four parameters from the lateral images
`for each lumbar spinal segment (Figs. 1, 2).
`The parameters measured included the distance between the
`lateral borders of the vertebral body in the plane of the upper end-
`plate, termed the upper vertebral width (UVW), and the distance
`between the anterior and posterior borders of the vertebral body,
`termed the upper vertebral depth (UVD). Similar measurements
`were made from the lower endplate, including the lower vertebral
`width (LVW) and lower vertebral depth (LVD). The distance be-
`tween the upper and lower endplates of the vertebral body at the
`posterior margin was measured from the lateral image and was
`termed the vertebral body height posterior (VBHp) and the anterior
`margin was termed the vertebral body height anterior (VBHa).
`Disc height (DH) was measured in the midline from the lateral im-
`age. The spinal canal width (SCW) was measured as the distance
`between the pedicles. Spinal canal depth (SCD) was defined as the
`distance from the posterior border of the vertebra to the lamina at
`the midline. Pedicle width (PDW) was also measured on the cross-
`sectional view of each vertebra. The pedicle height (PDH) was
`measured on the sagittal cut. Transverse process length (TPL) was
`the distance between the tips of the transverse processes measured
`
`measurement. Precision has varied considerably, particu-
`larly with respect to the imaging protocol and variables
`such as the magnification distance. Similarly, the size of
`study populations has frequently been limited, as has the
`number of samples studied.
`One large series was reported by Zindrick et al. [36],
`who studied 2905 vertebrae, although the number of pa-
`rameters studied was limited to the height, width, and
`transverse angles of the pedicles. Panjabi et al. [24]
`reported comprehensive studies of human cadaveric lum-
`bar vertebrae, but because of the extreme difficulty in ob-
`taining such specimens, the study was limited to only
`12 specimens. In addition, in cadaveric specimens it is dif-
`ficult to measure intervertebral disc height. Thus, compre-
`hensive measurements of vertebral and intervertebral di-
`mensions from a large series of samples have not been re-
`ported. An analysis of vertebral body circumference, the
`surface area of the vertebral endplates and the pedicle
`width has frequently been omitted from previous studies,
`and consequently there are limited data available on these
`characteristics [24, 29, 36]. Fang et al. published an impor-
`tant study in 1994 providing data applicable to the Asian
`lumbar spine, also obtained from CT scans, but these are
`not necessarily applicable to the Caucasian spine [11].
`Recently, developments in digitised images and ad-
`vances in computing have led to a new generation of dig-
`ital X-ray images, which permit image manipulation and
`enhancement. As a result, it is now possible to obtain
`measurements of the circumference and surface area of
`the endplate, an important consideration when designing
`implants for spinal fusion. These data permit the con-
`struction of anthropometric models for basic anatomical
`and biomechanical research and for pre-operative surgical
`preparation as well as for the design of spinal implants.
`The purpose of this study is to present data on the anthro-
`pometric characteristics of the lumbar vertebrae and as-
`pects of disc geometry from digitised CT images of the
`lumbar spine in a series of 126 patients.
`
`Materials and methods
`
`Study population
`
`This study was carried out on 126 patients presenting with low
`back pain and varying degrees of disc degenerative change to the
`Orthopaedic Spinal Clinic at the Hammersmith Hospitals NHS
`Trust between 1994 and 1996. There were 55 male patients, mean
`age 50 ± 13.60, and 71 female patients, mean age 49 ± 12.04 with
`an age range of 22–80 years. Patients with vertebral body abnor-
`malities, gross spinal pathology (including spondylolisthesis,
`retrolisthesis, disc space collapse) and those who had undergone
`spinal surgery were excluded.
`
`Measuring methods
`
`CT was performed using a Somatom Plus machine (Siemens) in
`the Department of Diagnostic Radiology. Sequential 3-mm contin-
`uous cross-sectional images were made parallel to both upper and
`
`Fig. 1 A lateral computed tomographic (CT) reconstruction with
`measurements on the fourth lumbar vertebra in a 47-year-old male
`subject (VBHp vertebral body height posterior, VBHa vertebral
`body height anterior, DH disc height, PDH pedicle height)
`
`2
`
`

`

`244
`
`from measurements of cross-sectional and lateral CT im-
`ages in 126 patients.
`
`Vertebral bodies
`
`The mean dimensions of the upper vertebral width was
`40.9 ± 3.6 mm in females and 46.1 ± 3.2 mm in males at
`L3, 46.7 ± 4.7 mm in females and 50.8 ± 3.7 mm in males
`at L4, and 50.4 ± 4.4 mm in females and 54.5 ± 4.9 mm in
`males at L5. The mean dimensions of the vertebral bodies
`for male spines were larger than for the female spines
`(P < 0.001). The depth and width of the vertebrae in-
`creased from L3 to L5 (P < 0.05). The anterior height of
`the vertebrae was the same for the third as for the fourth
`lumbar vertebrae (P < 0.05), but the posterior vertebral
`height decreased (P < 0.001).
`
`Spinal canal width and depth
`
`Figure 3 summarises data for the width and depth of the
`spinal canal. In the third lumbar vertebral body, the aver-
`age width was 24.2 mm and depth 16.1 ± 2.0 mm. For the
`fourth lumbar vertebral body, the mean canal width was
`23.6 ± 2.9 mm and depth 16.7 ± 2.7 mm. In the fifth lum-
`bar vertebral body, the mean canal width was 28.0 ± 3.9 mm
`and depth 17.1 ± 3.4 mm. There was no statistical differ-
`ence in spinal canal depth between male and female sub-
`jects (P > 0.05).
`
`Pedicle width and height
`
`Figure 4 summarises data for the width and height of
`the pedicles. At the L3 level, the pedicle width was 8.7 ±
`1.9 mm for females and 10.7 ± 2.0 mm for males. At the
`L4 level, it had increased to 11.3 ± 2.1 mm for females
`and 13.2 ± 2.0 mm for males. At the L5 level, the mean
`pedicle width was 15.3 ± 2.6 mm in females and 17.5 ±
`2.6 mm in males (P < 0.001). The pedicle height was
`14.1 ± 1.5 mm for females and 14.9 ± 1.6 mm for males
`at the L3 level, 13.9 ± 1.4 mm and 14.8 ± 1.6 mm at the
`L4 level and 13.4 ± 2.3 mm and 14.9 ± 1.8 mm at the L5
`level.
`
`Disc height
`
`There was no significant difference between the disc
`height at the L3/4 and L4/5 levels (P > 0.05). The L5/S1
`disc height was significantly less than at the L3/4 and
`L4/5 levels (P < 0.05). There was, however, considerable
`variation in disc height. The L4/5 disc height ranged from
`5.0 to 16.1 mm. Patients were subdivided according to
`disc height into four arbitrarily defined groups: 5.0–
`
`Fig. 2 A cross-sectional image of the fourth lumbar vertebral body
`in a 47-year-old male subject (UVW upper vertebral width, LVW
`lower vertebral width, UVD upper vertebral depth, LVD lower ver-
`tebral depth, SCW spinal canal width, SCD spinal canal depth,
`PDW pedicle width, TPL transverse process length, Cth cortical
`bone thickness)
`
`on the cross-sectional image. Cortical bone thickness (Cth) was as-
`sessed as the distance between the outer and inner borders of the
`lateral part of the vertebral body on the cross-sectional image. The
`level of the cross-sectional images at which the parameters were
`measured was 12 mm below the upper endplate. This level was se-
`lected to provide the clearest image to define all the necessary
`measurements in the average case.
`Within our series, the average disc height was 11 mm. CT im-
`ages from ten patients, five male and five female, with this disc
`height were selected for additional assessment of the cross-sec-
`tional area of the fourth lumbar vertebral body. The circumference
`and outline of the lower endplate was defined from the CT images
`by dividing the circumference into 5-mm segments with the cur-
`sor. The area of the endplate could be automatically calculated and
`was displayed in the information box.
`
`Repeatability of measurements
`
`To assess measurement errors, images of the fourth lumbar verte-
`bra from ten patients were randomly selected, and all parameters
`were measured on 2 consecutive days by the same observer. Data
`from the two sets of measurements were compared [2].
`
`Statistical analysis
`
`A statistical analysis was performed using the Stata statistical
`package (Stata Corporation, Texas, USA). A Student’s t-test was
`used to compare male and female data, and analysis of variance
`followed by orthogonal contrasts was used to compare the verte-
`bral dimensions at different spinal levels. A significance level of
`P < 0.05 was used. Repeatability was evaluated using Bland and
`Altman’s mean difference technique [2].
`
`Results
`
`Table 1 summarises the mean values, standard deviations
`and range of data for the lower lumbar spine obtained
`
`3
`
`

`

`Table 1 L3, L4, and L5 lum-
`bar vertebral body dimensions
`(mm) for 126 patients (mean ±
`SD) (UVW upper vertebral
`width, UVD upper vertebral
`depth, LVW lower vertebral
`width, LVD lower vertebral
`depth, VBHp vertebral body
`height posterior, VBHa verte-
`bral body height anterior, DH
`disc height, SCW spinal canal
`width, SCD spinal canal depth,
`PDW pedicle width, PDH
`pedicle height, TPL transverse
`process length, Cth cortical
`bone thickness)
`
`245
`
`Dimension
`
`Sex
`
`L3 and L3/4 disc
`
`L4 and L4/5 disc
`
`L5 and L5/S1 disc
`
`UVW
`
`UVD
`
`LVW
`
`LVD
`
`VBHp
`
`VBHa
`
`DH
`
`SCW
`
`SCD
`
`PDW
`
`PDH
`
`TPL
`
`Cth
`
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`M+F
`F
`M
`
`43.2 ± 4.3 (32.3–53.3)
`40.9 ± 3.6 (32.3–50.1)
`46.1 ± 3.2 (37.1–53.3)
`32.3 ± 3.3 (24.4–41.8)
`30.8 ± 3.1 (24.4–39.9)
`34.1 ± 2.6 (27.7–41.8)
`51.7 ± 4.8 (39.8–63.2)
`49.3 ± 4.1 (39.8–57.5)
`54.8 ± 3.6 (45.1–63.2)
`35.3 ± 3.6 (27.8–44.8)
`33.7 ± 3.1 (27.8–40.8)
`37.4 ± 3.1 (29.5–44.8)
`29.6 ± 2.4 (23.0–37.0)
`28.7 ± 2.2 (23.0–35.3)
`30.7 ± 2.1 (26.0–37.0)
`30.2 ± 2.1 (23.2–35.0)
`29.9 ± 2.3 (23.2–35.0)
`30.6 ± 1.8 (26.1–35.0)
`11.6 ± 1.8 (7.0–16.0)
`11.0 ± 1.6 (7.0–13.9)
`12.4 ± 1.7 (8.7–16.0)
`24.2 ± 3.1 (16.2–34.9)
`23.5 ± 2.3 (18.7–29.9)
`25.2 ± 3.6 (16.2–34.9)
`16.1 ± 2.0 (11.8–20.3)
`16.0 ± 2.1 (11.8–20.3)
`16.1 ± 1.9 (12.2–20.3)
`9.6 ± 2.2 (5.4–14.4)
`8.7 ± 1.9 (5.4–13.7)
`10.7 ± 2.0 (5.8–14.4)
`14.5 ± 1.6 (10.1–19.0)
`14.1 ± 1.5 (10.1–18.0)
`14.9 ± 1.6 (12.0–19.0)
`89.7 ± 9.2 (69.8–114.0)
`84.7 ± 6.7 (69.8–103.0)
`96.1 ± 8.0 (79.2–114.0)
`2.7 ± 0.4 (1.80–3.80)
`2.6 ± 0.4 (1.8–3.8)
`2.7 ± 0.4 (1.9–3.6)
`
`48.5 ± 4.7 (37.6–59.3)
`46.7 ± 4.7 (37.6–55.0)
`50.8 ± 3.7 (42.2–59.3)
`34.6 ± 3.6 (26.4–46.2)
`33.2 ± 3.3 (26.4–43.1)
`36.4 ± 3.2 (29.3–46.2)
`52.5 ± 4.7 (42.8–68.2)
`50.4 ± 4.2 (42.8–59.5)
`55.1 ± 4.1 (47.8–68.2)
`36.2 ± 3.7 (29.7–47.9)
`34.4 ± 2.8 (29.7–42.8)
`38.6 ± 3.4 (31.5–47.9)
`28.7 ± 2.3 (21.8–34.1)
`27.9 ± 2.3 (21.8–34.1)
`29.6 ± 1.9 (24.0–34.1)
`30.1 ± 2.4 (22.9–36.0)
`29.5 ± 2.4 (22.9–34.0)
`31.0 ± 2.1 (26.0–36.0)
`11.3 ± 2.1 (5.0–16.1)
`10.6 ± 2.0 (5.0–14.0)
`12.2 ± 2.0 (7.1–16.1)
`23.6 ± 2.9 (18.9–34.4)
`22.8 ± 2.5 (18.9–30.9)
`24.7 ± 3.2 (19.0–34.4)
`16.7 ± 2.7 (11.0–27.5)
`16.6 ± 2.7 (11.0–24.1)
`16.9 ± 2.8 (11.3–27.5)
`12.1 ± 2.2 (7.1–17.1)
`11.3 ± 2.1 (7.1–16.1)
`13.2 ± 2.0 (9.4–17.1)
`14.3 ± 1.5 (11.1–18.3)
`13.9 ± 1.4 (11.0–17.0)
`14.8 ± 1.6 (11.0–18.3)
`88.3 ± 9.1 (65.4–108.9)
`84.3 ± 7.8 (65.4–102.8)
`93.5 ± 7.9 (74.6–108.9)
`2.7 ± 0.4 (1.5–4.0)
`2.7 ± 0.4 (2.0–4.0)
`2.8 ± 0.4 (1.5–3.5)
`
`52.2 ± 5.1 (42.3–67.1)
`50.4 ± 4.4 (42.3–59.4)
`54.5 ± 4.9 (45.9–67.1)
`35.7 ± 3.7 (28.8–47.8)
`34.3 ± 3.5 (28.8–47.8)
`37.6 ± 3.1 (31.4–45.0)
`53.1 ± 6.0 (38.0–73.1)
`50.4 ± 4.9 (38.0–65.4)
`56.7 ± 5.3 (46.7–73.1)
`36.0 ± 4.0 (27.1–50.1)
`34.3 ± 3.3 (27.1–46.2)
`38.3 ± 3.8 (31.1–50.1)
`25.9 ± 2.0 (20.6–31.6)
`25.3 ± 1.9 (20.6–30.3)
`26.7 ± 1.9 (22.1–31.6)
`30.8 ± 2.5 (24.1–37.5)
`30.2 ± 2.6 (24.1–37.1)
`31.5 ± 2.1 (27.1–37.5)
`10.7 ± 2.1 (6.0–16.1)
`10.3 ± 2.1 (6.0–14.9)
`11.2 ± 2.0 (6.3–16.1)
`28.0 ± 3.9 (19.8–38.0)
`27.2 ± 3.6 (19.8–37.5)
`29.0 ± 4.0 (20.3–38.0)
`17.1 ± 3.4 (10.1–32.7)
`16.6 ± 3.1 (10.1–24.3)
`17.8 ± 3.7 (11.4–32.7)
`16.2 ± 2.8 (9.0–22.6)
`15.3 ± 2.6 (9.0–21.5)
`17.5 ± 2.6 (11.7–22.6)
`14.0 ± 2.2 (9.5–19.9)
`13.4 ± 2.3 (9.5–17.8)
`14.9 ± 1.8 (11.7–19.9)
`92.5 ± 8.4 (73.3–117.8)
`89.7 ± 7.2 (73.3–114.9)
`96.1 ± 8.6 (77.3–117.8)
`2.9 ± 0.5 (1.9–4.3)
`2.9 ± 0.5 (1.9–4.3)
`2.9 ± 0.5 (1.9–3.8)
`
`8.0 mm, 8.0–11.0 mm, 11.0–14.0 mm, 14.0–16.1 mm.
`These data provide information on the distribution of disc
`height in 126 patients and are illustrated in Fig. 5.
`
`Vertebral endplate surface area
`
`Table 2 presents the mean circumference and area of
`the fourth lumbar vertebral endplate in ten patients. The
`average circumference of the fourth lumbar vertebral end-
`plate was 141 ± 9.3 mm and the surface area was 1492 ±
`173.8 mm2.
`
`Intra-observer error
`
`Table 3 summarises the mean value, unit of the value,
`mean difference and the coefficient of repeatability of
`consecutive measurements in ten patients. In general, the
`limits of agreements were within 5% of the mean for most
`parameters [2].
`
`Discussion
`
`Measurements of human vertebrae have been performed
`by a number of authors [1, 7, 8, 11–13, 17, 23, 24, 26–28,
`
`4
`
`

`

`246
`
`Fig. 3 Spinal canal width and depth (mm) of the third, fourth and fifth lumbar vertebral bodies in males, females and both sexes com-
`bined. Error bars represent standard deviation
`
`Fig. 4 Pedicle width and height (mm) of the third, fourth and fifth
`lumbar vertebral bodies in males, females and both sexes com-
`bined. Error bars represent standard deviation
`
`Table 2 Measurement of the circumference and area (mm or
`mm2) of the fourth lumbar vertebral endplate in ten patients (C cir-
`cumference of the endplate)
`
`Patient
`
`Sex
`
`DH
`
`LVW
`
`LVD
`
`C
`
`F
`F
`F
`F
`F
`M
`M
`M
`M
`M
`
`A
`B
`C
`D
`E
`F
`G
`H
`I
`J
`Mean
`SD
`
`12.7
`12.3
`12.0
`12.2
`12.0
`12.2
`12.4
`12.0
`12.0
`13.0
`12.3
`0.3
`
`49.9
`48.2
`45.3
`45.6
`52.1
`53.8
`52.2
`54.8
`55.4
`55.3
`51.3
`3.8
`
`35.5
`35.8
`32.7
`32.6
`39.4
`38.1
`37.1
`36.0
`36.0
`40.5
`36.4
`2.6
`
`138.4
`134.9
`127.6
`125.2
`149.1
`147.7
`142.9
`147.2
`146.3
`151.7
`141.1
`9.2
`
`Surface
`area
`
`1430
`1412
`1223
`1199
`1664
`1651
`1517
`1579
`1566
`1679
`1492
`173.8
`
`equate and representative information, and a larger series
`such as that in the present study is required.
`In addition, the methods used in the past affect the ac-
`curacy of the information. It is, for example, difficult to
`obtain large numbers of cadaveric specimens, and also to
`provide appropriate information on disc dimensions from
`these specimens, which will have undergone post-mortem
`
`Fig. 5 The distribution of disc height (mm) in 126 male and fe-
`male patients
`
`33, 34]. The value of their data has depended on the num-
`ber of samples and the accuracy of measurement. In our
`study, the range for each parameter between the minimum
`and maximum was substantial. With such variation, as-
`sessment of a small number of samples cannot provide ad-
`
`5
`
`

`

`247
`
`Table 3 The mean value, mean difference and standard deviation
`(mm) of the difference for each variable as assessed by duplicate
`measurements in ten patients. The standard deviation of the differ-
`ence and mean value can be used to estimate the precision of each
`measurement (see text for details)
`
`Mean value
`
`Mean difference
`
`SD difference
`
`UVW
`UVD
`LVW
`LVD
`VBHp
`VBHa
`DH
`SCW
`SCD
`PDW
`PDH
`TPL
`Cth
`
`50.30
`34.80
`52.45
`35.84
`28.73
`30.27
`10.37
`22.64
`15.59
`12.15
`14.28
`84.72
`2.57
`
`0.12
`–0.06
`0.01
`–0.02
`0.71
`0.46
`0.28
`–0.10
`0.11
`–0.03
`0.14
`0.14
`–0.17
`
`0.26
`0.44
`0.39
`0.16
`0.99
`1.17
`1.07
`0.29
`0.41
`0.16
`1.33
`0.36
`0.26
`
`change. Early studies were carried out on plain X-ray
`films, but it is difficult to include an appropriate reference
`object in the focal plane, and errors are frequently intro-
`duced due to an inability to allow for the magnification
`factor.
`The introduction of CT provided the first real opportu-
`nity for appropriate assessment of cross-section, including
`vertebral body dimensions in living subjects. CT com-
`bined with the PACS measuring tool facilitates more ac-
`curate measurement, obtained with comparative ease, al-
`lowing a thorough assessment of a wide range of vertebral
`and intervertebral parameters in a larger number of pa-
`tients. The PACS instrumentation also permits manipula-
`tion of the CT data, with adjustment of contrast for opti-
`misation of image quality and measurement of distance,
`area and angle. Nevertheless, potential sources of error re-
`main. One source of error is the accurate identification of
`precise anatomical points. Intra-observer tests were car-
`ried out to analyse the magnitude of such errors. We found
`that the intra-observer error was in general less than 5%.
`Inter-observer error was not assessed, as all measurements
`for this database were performed by a single investigator.
`In the lumbar spine, the most common levels to be af-
`fected by significant abnormalities are the L3/4, L4/5 and
`L5/S1 discs. Intervertebral disc changes such as degener-
`ation with resorption or prolapse are common causes of
`low back pain. Unfortunately, there have been only a few
`previous reports on disc height in the lower lumbar verte-
`bral column, either from the normal population or from
`patients with low back pain. Saraste et al [28] reported the
`measurement of disc height on plain X-ray films, but it
`was confirmed in this paper that such techniques are too
`inaccurate for precise conclusions. Nevertheless, accurate
`knowledge of the dimensions of the disc space is crucial
`for studying low back pain and its causes. This informa-
`tion is important not only for basic research but also for
`
`clinical practice. Our study was carried out in patients
`with low back pain and may not represent appropriate val-
`ues for normal disc height in symptom-free individuals.
`However, the dose of irradiation associated with CT scan-
`ning is too great to permit studies of asymptomatic sub-
`jects, and the values obtained from patients with low back
`pain represent data from a population potentially liable to
`undergo spinal surgery, and thus provides data applicable
`for the design of spinal implants and surgical techniques.
`It was interesting to note the increasing pedicle width
`from L3 to L5. The safe insertion of pedicle screws de-
`pends on a sound and careful technique. The anatomical
`configuration is critical and, in particular, the dimensions
`of the screw to be inserted should be 80% or less of the
`outer diameter of the pedicle [32]. The decreasing pedicle
`size at L3 and L4 necessitates extreme care by the surgeon
`and, in most spinal units, pedicle screws are rarely used
`above L3 for degenerative lumbar spinal disease.
`Loss of disc height even in the absence of significant
`prolapse may lead to substantial and continued problems
`[5, 14, 35]. Bony encroachment on the neural foraminae
`leads to persistent root pain [16]. Techniques for inter-
`body spinal fusion have now been adapted to restore and
`maintain disc height [10, 30, 31], and various types of
`graft material and implant have been used for this purpose
`[3, 4, 15, 18, 19, 21, 22]. It is critical that the size is cor-
`rect. Too small an implant is liable to collapse into the
`centre of the vertebral body, but too large an implant
`makes surgical insertion more challenging and may lead
`to serious damage of surrounding structures. Closkey et
`al. [6] reported that the area covered with bone graft
`should be at least 30% of the total endplate in order to
`provide a margin of safety, whilst Pearcy et al. [25] con-
`cluded that at least 40% of the cross-sectional area should
`be covered by graft. If a restricted range of non-cus-
`tomised implants is to provide a satisfactory outcome in a
`full range of patients, it is essential for the designer and
`manufacturer of spinal implants to be aware of both the
`average and the range of endplate cross-sectional area.
`These data provide adequate information for the design
`of implants to treat patients with low back pain resulting
`from degenerative disease. CT, which inevitably involves
`exposure to a significant dose of radiation, is only justifi-
`able in symptomatic subjects who may require surgery.
`Stabilisation of patients with fractures involves the inser-
`tion of implants into those who were previously asympto-
`matic. In those previously well, there may be a greater av-
`erage disc height, but vertebral body dimensions should
`be little different from this series.
`In any event the cohort studied represented those most
`likely to require routine surgery in the average spinal unit.
`A substantial study of normal individuals can at present
`only be considered with magnetic resonance imaging
`(MRI), which is considered non-invasive. However this
`imaging technique is less accurate at defining the precise
`margins of osseous structures.
`
`6
`
`

`

`248
`
`These data from a large number of CT scans, coupled
`with accurate measurement with the PACS system, pro-
`vide the basis not only for anatomical studies and clinical
`research, but also for sensible rational implant develop-
`
`ment for a restricted inventory to promote a solution in the
`vast majority of cases. The evaluation of the potential ad-
`vantages of PACS in other situations will require further
`comparative studies.
`
`References
`
`1. Berry JL, Moran JM, Berg WS, Steffee
`AD (1987) A morphometric study of
`human lumbar and selected thoracic
`vertebrae. Spine 12: 362–367
`2. Bland M, Altman D (1986) Statistical
`measurements for assessing agreement
`between two methods of clinical mea-
`surement. Lancet 1:307–310
`3. Brantigan JW, Steffee AD (1993) A
`carbon fiber implant to aid interbody
`lumbar fusion. Spine 18: 2106–2117
`4. Calandruccio R A, Benton B F (1964)
`Anterior lumbar fusion. Clin Orthop
`35: 63–68
`5. Chen D, Fay LA, Lok J, Yuan P, Ed-
`wards WT, Yuan HA (1995) Increas-
`ing neuroforaminal volume by anterior
`interbody distraction in degenerative
`lumbar spine. Spine 20: 74–79
`6. Closkey RF, Parsons JR, Lee CK,
`Blacksin MF, Zimmerman MC (1993)
`Mechanics of interbody spinal fusion –
`analysis of critical bone graft area.
`Spine 18:1011–1015
`7. Eisenstein S (1977) The morphometry
`and pathological anatomy of the lum-
`bar spine in South African negroes and
`caucasoids with specific reference to
`spinal stenosis. J Bone Joint Surg Br
`59:173–180
`8. Eisenstein S (1983) Lumbar vertebral
`canal morphometry for computerised
`tomography in spinal stenosis. Spine
`8:187–191
`9. Elsberg CA, Dyke CC (1934) Diagno-
`sis and localization of tumours of
`spinal cord by means of measurements
`made on X-ray films of vertebrae and
`the correlations of clinical and X-ray
`findings. Bull Neurol Inst New York
`3:359–394
`10. Enker P, Steffee AD (1994) Interbody
`fusion and instrumentation. Clin Or-
`thop 300:90–101
`11. Fang D, Cheung K, Ruan D, Chan F
`(1994) Computed tomographic osteom-
`etry of the Asian lumbar spine.
`J Spinal Disord 7:307–316
`12. Gilad I, Nissan M (1985) Sagittal eval-
`uation of elemental geometrical dimen-
`sions of human vertebrae. J Anat 143:
`115–120
`13. Gilad I, Nissan M (1985) Sagittal radi-
`ographic measurements of the cervical
`and lumbar vertebrae in normal adults.
`Br J Radiol 58:1031–1034
`
`14. Hasegawa T, An HS, Haughton VM,
`Nowicki Bh, Wisconsin M (1995)
`Lumbar foraminal stenosis: critical
`height of the intervertebral discs and
`foramina. J Bone Joint Surg Am 77:
`32–38
`15. Kozak JA, Heilman AE, O’Brien JP
`(1994) Anterior lumbar fusion options.
`Clin Orthop 300:45–51
`16. Kuslich SD, Ulstrom CL (1991) The
`tissue origin of low back pain and sci-
`atica: a report of pain response to tis-
`sue stimulation during operations on
`the lumbar spine using local anaesthe-
`sia. Orthop Clin North Am 22: 181–
`187
`17. Larsen JL, Smith D (1980) Vertebral
`body size in lumbar spinal canal steno-
`sis. Acta Radiol Diagn 21:785–788
`18. Leong JC, Chow SP, Yau AC (1994)
`Titanium-mesh block replacement of
`the intervertebral disk. Clinical Orthop
`300:52–63
`19. Loguidice VA, Johnson RG, Guyer
`RD, Stith WJ (1988) Anterior lumbar
`interbody fusion. Spine 13:366–369
`20. LORAL medical imaging system
`(1995) Operating instructions: picture
`archiving and communication system.
`Publication no. 43 531 11; revision 01,
`8/66. Loral Medical Imaging Systems,
`Illinois
`21. Miyamoto K (1991) Long term follow-
`up results of anterior discectomy and
`interbody fusion for lumbar disc herni-
`ation. Nippon Seikeigeka Gakkai
`Zasshi 65:1179–1190
`22. Newman M, Grinstead G (1992) Ante-
`rior lumbar interbody fusion for inter-
`nal disc disruption. Spine 17:831–833
`23. Nissan M, Gilad I (1984) The cervical
`and lumbar vertebrae – an anthropo-
`metric model. Eng Med 13:111–114
`24. Panjabi MM, Goel V, Oxland T,
`Takata K, Duranceau J, Krag M, Price
`M (1992) Human lumbar vertebrae
`quantitative three-dimensional anato-
`my. Spine 17:299–306
`25. Pearcy MJ, Evans JH, O’Brien JP
`(1983) The load bearing capacity of
`vertebral cancellous bone in interbody
`fusion of the lumbar spine. Eng Med
`12:183–184
`26. Postacchini F, Pezzeri G, Montanaro
`A, Natali G (1980) Computerised to-
`mography in lumbar stenosis – a pre-
`liminary report. J Bone Joint Surg Br
`62:78–82
`
`27. Postacchini F, Ripani M, Carpano S
`(1983) Morphometry of the lumbar
`vertebrae – an anatomic study in two
`caucasoid ethnic groups. Clin Orthop
`172:296–303
`28. Saraste H, Brostrom LA, Aparisi T,
`Axdorph G (1985) Radiographic mea-
`surement of the lumbar spine – a clini-
`cal and experimental study on man.
`Spine 10:236–241
`29. Scoles P, Linton A, Latimer B, Levy
`M, Digiovanni B (1988) Vertebral
`body and posterior element morphol-
`ogy: the normal spine in middle life.
`Spine 13: 1082–1086
`30. Steffee AD (1989) The variable screw
`placement system with posterior lum-
`bar interbody fusion. In: Lin PM, Gill,
`K (eds) Lumbar interbody fusion: prin-
`ciples and techniques in spine surgery.
`Aspen, Rockville, p81–93
`31. Steffee AD, Sitkowski DJ (1988) Pos-
`terior lumbar interbody fusion and
`plates. Clin Orthop 227: 99–102
`32. Suk SI, Cha SI, Lee CK, Kim WJ
`(1995) A study on the pullout strength
`of pedicle screws in relation to the size
`of the drill holes and the pedicle
`screws. Presented at 30th Annual
`Meeting of the Scoliosis Research So-
`ciety, Ashville, North Carolina, Sep-
`tember 13–16
`33. Twomey LT, Taylor R (1987) Age
`changes in lumbar vertebrae and inter-
`vertebral discs. Clinical Orthop 224:
`97–104
`34. Van Schaik JJ, Verbiest H, Van Schaik
`FD (1985) Morphometry of lower lum-
`bar vertebrae as seen on CT scans:
`newly recognized characteristics. AJR
`145:327–335
`35. Vuono-Hawkins M, Langrana NA, Par-
`sons JR, Lee CK, Zimmerman MC
`(1995) Materials and design concepts
`for an intervertebral disc spacer. II.
`Multidurometer composite design.
`J Appl Biomater 6:117–123
`36. Zindrick M, Wilts L, Doornik A,
`Widell E, Knight G (1987) Analysis of
`the morphometric characteristics of the
`thoracic and lumbar pedicles. Spine 12:
`160–166
`
`7
`
`