|Home | About | Journals | Submit | Contact Us | Français|
Previous studies have reported position-dependent changes of the lumbar intervertebral foramen (LIVF) dimensions at different static flexion-extension postures. However, the changes of the LIVF dimensions during dynamic body motion have not been reported.
The objective of this study was to investigate the in-vivo dimensions of the LIVF during a dynamic weight-lifting activity.
A retrospective study.
Ten asymptomatic subjects were recruited for this study. 3D vertebral models of the lumbar segments from L2 to S1 were constructed for each subject using MR images. The lumbar spine was then imaged using a dual fluoroscopic imaging system as the subject performed a dynamic weight-lifting activity from an upper body position of 45° to a maximal extension position. The in-vivo positions of the vertebrae along the motion path were reproduced using the 3D vertebral models and the fluoroscopic images. The minimal area, height, and width of each LIVF during the dynamic body motion were analyzed.
The LIVF area and width monotonically decreased with lumbar extension at all levels except L5-S1 (p < 0.05). On average, the LIVF area decreased by 7.4 ± 6.7 %, 10.8 ± 7.7 % and 10.0 ± 8.0 % at the L2–3, L3–4 and L4–5 levels, respectively, from the flexion to the upright standing position, and by 6.4 ± 5.0 %, 7.7 ± 7.4 % and 5.1 ± 5.1 %, respectively, from the upright standing to the extension position. LIVF height remained relatively constant at all segments during the dynamic activity. The foramen area, height, and width of the L5-S1 remained relatively constant throughout the activity.
Human lumbar foramen dimensions show segment-dependent characteristics during the dynamic weight-lifting activity.
Lumbar intervertebral foramen (LIVF) stenosis is a common cause of nerve root compression with the symptom of radiculopathy. This condition often requires surgical decompression or implantation of an interspinous process device if conservative management fails [1–3]. The boundary of a LIVF consists of movable facet joints. Therefore, the geometry of the LIVF changes with lumbar positions during daily activities, potentially inducing mechanical stimulus on pain-sensitive structures. For example, in the extended position, LIVF could cause increased compression on the nerve root from facet motion and ligamentum flavum bulging [4, 5]. Any complication after decompression could result in failed back surgery syndrome [6, 7]. Accurate knowledge on the geometric changes of the LIVF is crucial for the diagnosis and treatment of LIVF stenosis. Therefore, numerous studies have been carried out to investigate changes in LIVF geometry using in-vitro and in-vivo experimental setups [8–16].
Previous in-vitro studies have described a decrease in the LIVF dimensions using lumbar cadaveric specimens moving from flexion to extension [8–11]. In-vivo studies have indicated position-dependent changes of the LIVF dimensions at different static flexion-extension postures using open-MRI techniques [14, 16]. Although these studies have dramatically improved our knowledge of LIVF dimensions, the in-vivo changes during dynamic functional flexion and extension are still unclear. Further, few data has been reported on the differences in foramen changes between segment levels.
The purpose of this study was to investigate the in-vivo dynamic changes of the LIVF dimensions during a weight-lifting activity from a flexion position of 45° to a maximal extension position. A validated combined dual fluoroscopic imaging system (DFIS) and MR-based 3D vertebral modeling method was used to determine the dynamic changes of the LIVF dimensions in the lumbar spine . The LIVF dimensions at different segment levels were specifically compared during the weight-lifting activity.
Ten asymptomatic subjects (5 females and 5 males, aged from 40 to 60 years old) were recruited with approval of the authors’ institutional review board (IRB). Exclusion criteria included: current or prior back pain, anatomic abnormalities, or any spinal disorders. Written consent was obtained from each subject prior to the experiment.
Each subject was scanned in a supine, relaxed position using a 3.0 Tesla scanner (MAGNETOM Trio, Siemens, Germany) with a spine surface coil and a T2-weighted fat-suppressed 3D SPGR sequence. Parallel digital images of the lumbar spine were acquired with a 1.0 mm thickness, no gap, and a resolution of 512 × 512 pixels. 3D vertebral models of the lumbar segments from L2 to S1 were constructed for each subject using the MR images in a solid modeling software program (Rhinoceros®, Robert McNeel & Associates, Seattle, WA) . Figure 1 shows a typical 3D model of the lumbar segments from L2 to S1.
Subsequently, the lumbar spine was imaged using the DFIS (BV Pulsera, Phillips, Bothell, WA) as the subject performed lumbar extension, from a flexion position of 45° to a maximal extension position, with each hand holding an 8 pound dumbbell  (Fig. 1). Pelvis motion was limited during the weight-lifting activity by a custom-made frame. During the fluoroscopic imaging, the regions above and below the subject’s lumbar spine were protected using specifically designed vests, skirts, and thyroid shields to minimize the radiation exposure. The fluoroscopy captured the dynamic positions of the vertebrae at 30 frames per second with an 8 milli-second pulse width. The motion took approximately two seconds, resulting in about 60 pairs of fluoroscopic images being collected.
The in vivo positions of the vertebrae along the dynamic motion path of the weight-lifting activity were reproduced in the Rhinoceros® software using the 3D vertebral models and the fluoroscopic images (Figs. 1 and and2).2). This was done by recreating the geometry of the DFIS in the software, and rotating and translating each vertebrae in six-degrees-of-freedom until their positions best matched the fluoroscopic images [17, 20–22]. The accuracy of this technique has been shown to be within 0.3 mm and 0.7° for determining dynamic vertebral positions and orientations, respectively .
After reproducing the in-vivo vertebral positions, the 3D LIVF models were obtained using the Boolean operator in the solid modeling software (Figs. 3A and 3B). The minimal cross-sectional area of each LIVF was determined at the pedicle cutting plane using the 3D LIVF model (Fig. 3B). The height of the LIVF was defined as the longest distance between the cranio-caudal boundary (green line) (Fig. 3C) . The width of the LIVF was defined as the shortest distance between the postero-inferior corner of the proximal vertebrae and the opposing boundary (blue line) (Fig. 3C) . The area of the LIVF was drawn anatomically according to the LIVF bony outline (red outline) (Fig. 3C) . The LIVF dimensions of each intervertebral level were averaged from both sides of the foramen.
In this study, we investigated the LIVF dimensions (area, height, and width) of L2 to S1 vertebral levels during the weight-lifting activity. Three positions along the motion path were selected for analysis: 45° flexion, upright position, maximal extension. A repeated measures analysis of variance (ANOVA) was used to compare the LIVF dimensions along the motion path and across the lumbar vertebral levels. A statistical difference was achieved when p-value < 0.05. A Newman-Keuls post-hoc test was performed when a statistical difference was detected. Statistical analysis was performed using Statistica software (Statsoft, Tulsa, OK, USA).
During the weight-lifting activity, the range of motion of the L2 to S1 segment was 29.8 ± 8.2 °. Each vertebral level showed similar ranges of motion (p > 0.05) (Table 1).
The LIVF area and width decreased significantly at all levels except L5-S1 along the lumbar motion path (p < 0.05) (Figs. 4 and and5).5). From 45° flexion to the upright position, LIVF area decreased by 7.4 ± 6.7 % at L2–3, 10.8 ± 7.7 % at L3–4, and 10.0 ± 8.0 % at L4–5, with almost no change observed at L5-S1 (0.0 ± 6.0 %) (Table 1). From the upright to extension positions, the LIVF area decreased by 6.4 ± 5.0 % at L2–3, 7.7 ± 7.4 % at L3–4, and 5.1 ± 5.1 % at L4–5, with minimal change again observed at L5-S1 (1.7 ± 10.5 %) (Table 1). Similar trends were also seen for LIVF widths (Fig. 5). However, foramen heights did not significantly change with body position at any level. (Fig. 6)
In general, higher segment-levels showed larger foramen area, especially in the flexion position (Fig. 4). For example, at the 45° flexion position, the LIVF area of the L4–5 was 179.2 ± 25.0 mm2, which was significantly larger than that of L5-S1 (149.7 ± 19.4 mm2) (p < 0.05). In the upright position, the LIVF area of L4–5 was 163.3 ± 23.1 mm2, which was also larger than that of L5-S1 (149.9 ± 19.6 mm2). In the maximal extension position, LIVF area of the L4–5 and L5-S1 were similar (155.2 ± 25.1 mm2 and 148.1 ± 27.4 mm2, respectively) (Fig. 4).
This study investigated the in-vivo dynamic changes of the LIVF dimensions during a functional extension activity of the upper body. The data indicated that the LIVF area decreased during the weight-lifting activity at levels L2–3, L3–4, and L4–5. The decreasing areas were similar in trend to the decreases in LIVF width, while the heights of the LIVFs were relatively constant throughout the activity. Interestingly, L5-S1 demonstrated distinct characteristics from the other levels, as the foramen area, height, and width remained relatively constant throughout the activity. The LIVF dimension changes did not correspond the ranges of extension of the vertebral levels during the lumbar activity.
Several studies have measured the dimensional changes of the LIVF during the flexion-extension movement using human cadaveric specimens. Panjabi et al  found that the cross-sectional area of the neural foramen increased by 30% with flexion and decreased by 20% with extension in the non-degenerated spine. Mayoux-Benhamou et al  showed a reduction of 17.9% in LIVF height and 14.1% in LIVF width as the cadaveric spine moved from flexion to extension. Inufusa et al  and Fujiwara et al  reported reductions in LIVF area of 24% and 21%, respectively, as the spine moved from flexion to extension.
Some in-vivo studies of LIVF dimensions have been performed using CT or open-MRI techniques. Iwata et al  used CT imaging to study 12 asymptomatic young volunteers while applying axial loads using a compression device. They found that the in-vivo LIVF dimensions decreased significantly at all levels, but increased at the L5-S1 during the axial loading. Zamani et al  used an open-MRI technique to investigate the LIVF area in a sitting position while performing flexion and extension. They showed LIVF area decreased in extension and tended to increase in flexion. Schmid et al  evaluated the physiologic changes of LIVF area in upright neutral, upright flexed, upright extended, and supine extended positions using an open-MRI technique, and observed position-dependent changes in the LIVF area. Recently, similar conclusions on position-dependent changes in the LIVF area were also reported by Singh .
In general, these previous in-vitro and in-vivo results have shown similar trends in foramen area changes during the flexion and extension motion as our study. However, it is difficult to make a quantitative comparison among these studies due to the various technologies used. Our data showed an area change of about 10% from 45° flexion to upright standing, and from upright s tanding to maximal extension. The in-vitro studies applied loading conditions that may be different from in-vivo physiological loading. In-vivo static loading conditions could also be different from in-vivo dynamic loading conditions. Further, it should be noted that we reconstructed 3D models of the LIVFs and accurately defined the cross sections of the minimal foramen areas. However, the mid-sagittal plane of the cranio-caudal pedicles (mid-sagittal plane) [10–12] or the plane along the middle of the cranio-caudal pedicles (mid-pedicle plane)  has been used in the literature to define the narrowest foramen areas (Fig. 7). These different cutting planes could also lead to variations in measurements of the foramen areas. Table 2 compares the foramen area data of the ten subjects measured using three different cutting methods. Substantial differences in foramen areas were seen if different cutting planes were used for foramen area measurements.
For patients with LIVF stenosis, nerve root impingement is typically exacerbated in extension and relieved in flexion [10, 25]. Penning et al  used CT myelography to investigate symptomatic patients in flexion and extension positions and reported that the posture-dependent bilateral nerve root involvement was common at the L3–4 and L4–5 levels, but less common at L5-S1. At the L5-S1 level, the L5 nerve root is larger in diameter, but the LIVF area is smaller than other lumbar levels . However, the higher root/foramen area ratio does not enhance the nerve root involvement at the L5-S1 level during flexion-extension motion [12, 15, 16, 26]. This may be due to the less variation of the structural dimensions of the L5-S1 LIVF along the flexion-extension path. The distinct characteristics of the L5-S1 level foramen geometry, which showed no decrease in area throughout the dynamic motion, might therefore be a protective mechanism for nerve root involvement during the flexion-extension . Our data indicated that the LIVF area is more affected by the width change than by the height change of the foramen. Therefore, it may be clinically important to consider the geometric features of the foramen in width and height directions when performing a decompression surgery. Interestingly, the range of motion of the L5-S1 level is similar to the other segmental levels during the flexion-extension motion. Further research is warranted to determine the exact biomechanical factors that lead to the varied geometric characters of the LIVFs of the human body, since an accurate knowledge of the dynamic changes of LIVF dimensions may provide implications for accurate diagnoses and improved treatments of lumbar foramen stenosis.
Several limitations to this study should be considered. Firstly, a relatively small number of subjects were investigated. Many factors could influence the anatomic structures of the LIVF shape and size, including: age; gender; level; vertebral and pedicle geometries; and the orientation of lumbar facet joints [23, 27–30]. Future studies should sub-divide the subjects into smaller groups to include variables such as age, gender, and body height/weight. Our study also used 3D vertebral models without soft tissues due to the difficulties of specifying soft tissue boundaries around the foramen area as well as the deformation of the soft tissues during dynamic lumbar motion. Therefore, the posterior disc and ligaments were not considered. Future studies should take these effects into consideration with technology advancement. Despite the above limitations, this study represents the first report of dynamic in-vivo changes in LIVF dimensions, and reveals distinct characteristics of the L5-S1 LIVF.
In conclusion, this study investigated the in-vivo dynamic changes in cross-sectional measurements of the LIVFs during a functional activity. We found that the LIVF area and width decreased significantly at all levels except L5-S1 during lumbar extension. The L5-S1 demonstrated distinct characteristics from other levels as foramen area, height, and width remained relatively constant throughout the activity. The results provide further insight into in-vivo dynamic function of the lumbar spine. The data may also provide implications for improvement of diagnoses and treatments of lumbar foramen stenosis.
The authors would like to gratefully acknowledge the financial support from the National Institute of Health (R21AR057989), Synthes, Inc., Department of Orthopaedic Surgery of Massachusetts General Hospital and China Scholarship Council (201306370122).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of interest