Physical exercise is widely promoted for apparently healthy individuals, and evidence is growing to support exercise training interventions for patients with chronic disease. Current American College of Sports Medicine guidelines state that most adults should engage in moderate-intensity cardiorespiratory exercise training for ≥30 min·d−1 on ≥5 d·wk−1, vigorous-intensity cardiorespiratory exercise training for ≥20 min·d−1 on ≥3 d·wk−1 (≥75 min·wk−1), or a combination of moderate- and vigorous-intensity exercise to achieve a total energy expenditure of ≥500 to 1000 MET·min·wk−1 (15), where MET represents the ratio of the rate of energy expended during an activity to the rate of energy expended at rest. In addition, weight-bearing exercise modalities, such as walking and running, are often advocated to maintain spinal health.
The two main types of spinal loading during ambulation are compressive loading and impact loading. Compressive loading is caused by weight-bearing activities (6), and impact loading is caused by the landing phase of impact activities such as running, where impacts are transmitted from the limbs to the spine through repeated landings (23). Running inflicts a ground reaction force of two to three times the body weight that lasts for 2–3 s (8) and is transmitted to the spine (24). In addition, running involves repeated torso rotations because of the movement of the arms, which could exert additional load on the thoracic region of the spine.
Precision stadiometry has been demonstrated to be a reliable method of measuring the change in total stature (13). Throughout the course of a normal day, healthy individuals experience a loss in total stature of approximately 1% (25), where serial measurements demonstrate that the rate of stature loss peaks shortly after rising (4). Loss in stature is assumed to occur when osmotic pressure in the discal tissues is exceeded by the load on the spine, which results in fluid being expelled from the intervertebral disc (IVD) (20). Loss in stature has been quantified during a wide variety of occupational and recreational contexts (2,3,10), including ambulatory activities (19,26). However, loss in stature provides an overall measurement of the change in stature rather than changes within spinal structures; consequently, this measurement technique neither accounts for changes in spinal curves (e.g., lordosis and kyphosis) or deformation in other tissues (e.g., hips, knees, ankles, and heel pads) nor quantifies changes in structures within different regions of the spine.
Magnetic resonance imaging (MRI) can be used to quantify the size of IVDs without exposing the participant to x-rays or other forms of harmful radiation. Han et al. (17) digitized key landmarks on IVDs from MRI to study the diurnal changes in disc height variation in rats. Violas et al. (27,28) demonstrated that digitization and custom-made image processing software can be used to quantify the volume of lumbar IVDs in human patients with idiopathic scoliosis. Midsagittal MRI analyses using similar procedures have demonstrated that static loading during sitting (14) and walking (21) causes reductions in IVD height in the lumbar vertebral region. Recently, 1 h of running was demonstrated to reduce IVD height in lumbar discs of long-distance running athletes (11). In contrast, the influence of a recommended bout of moderate-intensity running on IVD height and volume remains unknown. Furthermore, the susceptibility of IVDs in the lumbar and thoracic regions to compression during exercise has not been quantified in healthy adults.
Therefore, the aim of this investigation was to use repeated MRI examinations to quantify the acute effects of a 30-min bout of moderate-intensity treadmill running on IVDs throughout the thoracic and lumbar vertebral regions.
This study used a single-group repeated-measures design. Participants attended three testing sessions on different days, separated by 3–8 d. Eight male volunteers (18–23 yr of age) with no history of disease or back pain were informed about the potential risks of the study and gave written informed consent to participate in the study, which was approved by a university ethics committee. The experimental procedures were in accordance with the policy statement of the American College of Sports Medicine.
The first session was used to take preliminary measurements, to familiarize the participants with anthropometry and treadmill running, and to calculate the exercise intensity required for the exercise trial. During the second and third sessions, the participants reported to the clinic within 30 min of rising from bed, after at least 10 h of bed rest, and having undertaken minimal ambulatory activity. During the second session, venous blood samples were taken and baseline measurements of body mass and stature obtained before spinal MRI examination. During the third session, the same measurements were repeated on arrival (preexercise) and after completing 30 min of treadmill running at 70% HRreserve (postexercise).
Participants were asked to refrain from exercise within 24 h of testing, refrain from consuming food for a minimum of 10 h before testing, and abstain from consuming substances that affect normal hydration status (e.g., diuretics and alcohol) within 24 h of testing.
During the preliminary testing session, the participants completed baseline measurements of stature (Portable Stadiometer; Holtain, UK) and body mass (model 712; seca, Germany), followed by a progressive treadmill run to exhaustion. The treadmill run was used to familiarize the participants with exercise and to calculate the absolute running speed required to elicit an exercise intensity of 70% HRreserve, which was used to determine the running speed during the third session.
Standing stature was measured according to the International Society for the Advancement of Kinanthropometry guidelines (22) using a portable stadiometer (Holtain). Briefly, participants stood on the stadiometer for at least 2 min to allow for heel pad compression before their heads were adjusted to the Frankfurt plane. The caliper of the stadiometer was lowered so that it rested on the vertex of the head. Subjects were instructed to take a normal inhalation extending their spines as much as they could, and the measure was taken. Each measure was taken three times, and the average value was used for analysis. The total time taken to complete this procedure was less than 4 min. The anthropometrist was blinded to previous measurements. Exercise-induced loss of stature was calculated as the difference in preexercise to postexercise stature.
Blood sampling and analyses.
Baseline and preexercise venous blood samples were taken from an antecubital vein (Vacutainer system; Becton-Dickinson Ltd., UK) and collected in 5-mL containers (Becton-Dickinson Ltd.) with the anticoagulant ethylenediaminetetraacetic acid. After centrifuging at 3000g for 15 min, plasma osmolality was measured in duplicate by freezing point depression using a cryoscopic osmometer (Osmomat 030; Gonotec, Berlin, Germany) to determine hydration status.
Progressive treadmill run.
After completion of a 5-min warm-up on a treadmill (STEX 7520T; Afton, Richmond, VA) inclined at an angle of 10° at 2.0 mph followed by a 5-min passive rest period, the test began at 2.0 mph and increased in speeds every 2 min until volitional exhaustion. Heart rate (S810; Polar, Finland) and RPE (5) were monitored throughout the test. This test was used to measure peak HR, estimate V˙O2max, and determine a running speed required to elicit a relative exercise intensity of 70% HRreserve.
Three MRI examinations (baseline, preexercise, and postexercise) were carried out in the supine position, lasting between 7 and 10 min. The MRI protocol was performed on a 1.5-T high-definition 16-channel system (GE Medical Systems, Waukesha, WI). Sagittal T2 fast relaxation fast spin echo sequences were used to image the cervical, dorsal, and lumbar regions after a 3-plane localizer. The time to repeat was 2200–3000 ms and effective echo time of 110 ms. Slice thickness was 3.2 mm. The field of view was 71 cm for the sagittal images with an image matrix of 352 × 320 and a number of excitations equal to 4. The cervical, dorsal, and lumbar images were obtained in separate sections and subsequently fused using the MRI pasting software on the workstation (Advantage Windows, GE Healthcare, Waukesha, WI). All images were stored in DICOM format, exported as uncompressed full-size images (Centricity; GE Healthcare), and imported into LabVIEW for digitization (Professional version 10.0; National Instruments, Austin, TX).
The margins of the vertebral bodies from T1 to S1 were digitized for all images where the vertebral endplate and associated IVD were visible. At least seven points were digitized along the superior and inferior vertebral endplates (Fig. 1A). These digitized points were interpolated in 1-mm intervals, and these coordinates were used to determine the distance between adjacent vertebral endplates (Fig. 1B). Digitization was performed by a single operator after extensive training and familiarization. All images derived from the MRI scan were combined to produce a digital three-dimensional representation of all thoracic and lumbar IVDs to determine mean vertical IVD height and to calculate IVD volume (Fig. 1C). Intraobserver reliability was determined using 10 repeated measurements of a randomly selected example for all IVD locations. Repeated measurements were completed on separate days, and the operator was blinded to previous measurements. Intraclass correlation and standard error of measurement for IVD height and volume were 0.99 (95% confidence interval (CI), 0.98–1.00) and 0.027 mm (95% CI, 0.023–0.030 mm) and 0.99 (95% CI, 0.99–1.00) and 0.07 mm3 (95% CI, 0.06–0.08 mm3), respectively.
Statistical analysis was carried out using Predictive Analytics Software Statistics (version 18.0; SPSS Inc., Chicago, IL). Shapiro–Wilk tests were conducted to assess the normality of all data. Group data were expressed as mean ± SD, and statistical significance was set at P < 0.05. Two-way repeated-measure ANOVAs (within subject: timing and IVD) were used to determine the influence of day-to-day variation and exercise (timing: baseline, preexercise, and postexercise) and IVD (IVD: 17 IVD locations) on IVD height and volume. Significant main effects of timing and IVD were further investigated using pairwise comparisons with Bonferroni adjustment. One-way repeated-measure ANOVAs were used to evaluate the influence of timing on plasma osmolality and stature. Significant main effects were followed up using pairwise comparisons with Bonferroni adjustment. Bonferroni adjustments were included in the calculations for the 95% CIs for differences between means.
Participants arrived at the clinical facility on the second and third days with plasma osmolality being 289 ± 1 and 285 ± 2 mOsm·kg−1, respectively. Participant characteristics were as follows: age, 21.1 ± 0.8 yr; mass, 69.4 ± 5.1 kg; stature, 174 ± 2 cm; and estimated V˙O2max, 42.6 ± 0.6 mL·kg−1·min−1.
Changes in MRI measurements.
Figure 2 displays mean IVD heights throughout the thoracic (T1–T2 to T12–L1) and lumbar region (L1–L2 to L5–S1) during the three MRI scans (baseline, preexercise, and postexercise). Mean IVD heights were different between MRI scans (main effect time: F2,14 = 74.1, P < 0.001, ηp2 = 0.91). Baseline and preexercise values were similar with day-to-day variations of 0.04 mm (95% CI, −0.04 to 0.12 mm) or 0.6% ± 0.6% (P = 0.706); however, preexercise to postexercise mean IVD height reduced by 0.33 mm (95% CI, 0.27–0.39 mm) or 6.3% ± 0.9% of preexercise values (P < 0.001). Mean IVD height in the thoracic and lumbar regions are presented in Figure 3, where day-to-day variations were 0.03 ± 0.07 and 0.07 ± 0.20 mm and exercise-induced changes were 0.34 ± 0.09 and 0.29 ± 0.09 mm, respectively. Relative losses in exercise-induced changes in mean IVD height were greater in the thoracic region (7.7% ± 1.2%) when compared with the lumbar region (3.7% ± 0.6%, P < 0.001). The total exercise-induced loss in IVD height within the thoracic (12 IVDs) and lumbar region (5 IVDs) was 4.12 ± 1.10 and 1.45 ± 0.45 mm, respectively.
Calculated IVD volumes were different between MRI scans (F2,14 = 443.8, P < 0.001, ηp2 = 0.98; Fig. 3). Post hoc analyses revealed that the day-to-day variation in estimated IVD volume was 0.04 mm3 (95% CI, −0.06 to 0.15 mm3) or 0.4% ± 0.6% (P = 0.706), whereas exercise-induced reductions in mean IVD volume was 754 mm3 (95% CI, 643–866 mm3) or 6.9% ± 1.0% of preexercise values (P < 0.001). Although exercise-induced losses in total IVD volume was greater in the lumbar region (3122 mm3; 95% CI, 2649–3594 mm3) compared with the thoracic region (2856 mm3; 95% CI, 2413–3299 mm3) (Fig. 4), the relative change in IVD volume in the thoracic and lumbar regions were 8.4% ± 1.4% and 4.2% ± 0.6%, respectively.
Changes in stature.
Stature differed across measurements (main effect time: F2,14 = 146.9, P < 0.001, ηp2 = 0.96). Baseline and preexercise measurements were similar with mean differences of 0.4 mm (95% CI, −0.4 to 1.2 mm; P = 0.591). Exercise-induced changes in stature equated to a shrinkage of 8.1 ± 0.6 mm (95% CI, 6.1–10.1 mm; P < 0.001).
The main findings from this study were that 30 min of moderate-intensity treadmill running resulted in a 6.3% ± 0.9% reduction in mean IVD height and a 6.9% ± 1.0% reduction in mean IVD volume. Resting mean IVD height and volume were similar across days (P ≥ 0.706) with day-to-day variation being 0.6% ± 0.6% and 0.4% ± 0.6%, respectively. Furthermore, exercise-induced changes in IVD height and volume were evident in IVDs in the thoracic and lumbar regions. This is the first study to demonstrate that IVD height and volume are reduced throughout the thoracic and lumbar regions of the spine after an acute bout of running exercise.
The current study demonstrated that 30 min of treadmill running at 70% HRreserve resulted in an 8.1 ± 0.6-mm loss of total stature, which is similar to the amount of stature loss (7.7 ± 1.0 mm) previously reported after marathon runners completed 30 min of treadmill running at 100% of race pace (16). The magnitude of loss in stature is significant when compared with usual diurnal changes of 15–19 mm (24). In the present study, stature measurements were taken using the International Society for the Advancement of Kinanthropometry measurement techniques with a commercially manufactured precision stadiometer. Nevertheless, the day-to-day variation in stature measurement was relatively small (95% CI, −0.4 to 1.2 mm), which confirms the suitability of the measurement technique used in the present study.
Although the majority of authors have generally attributed the running-induced loss in total stature to changes in IVD height and volume (1,12,16,19), this assumption had not previously been directly evaluated. We have demonstrated, using seated x-ray images, that treadmill running at 70% HRreserve caused a loss of vertical height in the lumbar vertebral region of 4.4 ± 0.8 mm (18). Direct measurements from x-ray images provide good contrast between the vertebral endplate and connecting tissues; however, the lack of depth perception makes this imaging method impractical to quantify mean IVD height and volume. In contrast, MRI provides a three-dimensional reconstruction of IVDs using a nonradiant imaging technique that has been shown to provide reproducible and reliable volume measurements for lumbar IVDs (27).
Mean IVD height was determined for IVDs in the thoracic (T1–T2 to T12–L1) and lumbar (L1–L2 to L5–S1) regions (Fig. 2). The values for mean IVD height in this study are similar to the lumbar IVD heights calculated using computed tomography imaging (30); however, it should be noted that these values are not directly comparable because the mean IVD height in the current study provides a measure of central tendency in IVD height across the three-dimensional image of the IVD, whereas Zhou et al. quantified IVD height at a single measurement at the middle of the IVD taken from the midline of the lateral computed tomography image.
The mean loss in IVD height calculated in the current study (0.33 mm; 95% CI, 0.27–0.39 mm) was less than the reported reduction in mean lumbar IVD height (0.66 to 1.10 mm) after athletes completed 1 h of running (11). In contrast to our study, Dimitriadis et al. (11) reported change in the mean value of anterior and posterior IVD height from midsagittal MRIs obtained while participants were seated in an upright position. It is likely, therefore, that differences in measurement technique, compressive loading, and dose of running explain differences in the magnitude of change in mean IVD height.
Although absolute exercise-induced changes in IVD height and volume were similar in the thoracic and lumbar spinal regions, relative losses in IVD height and volume in the thoracic region were approximately twice the respective values calculated for the lumbar region. It is possible that rotations of the thorax during running develop large compressive forces on IVDs in the thoracic region. However, small axial rotations of in vitro human IVDs do not influence intradiscal pressures or disc height (29); consequently, this conclusion is speculative. Alternatively, it is possible that IVDs in the lumbar region are better adapted to resist compression or recovered more rapidly during the supine MRI procedure than thoracic IVDs. The supine MRI allowed repeatable IVD measurement while the spine was under minimal compressive load; however, it is likely that some IVD recovery would have occurred during the MRI procedure. In either case, these findings demonstrate that both regions of the vertebral column are subject to loss in IVD height and volume; consequently, future studies that aim to investigate the influence of physical activities and/or interventions on IVD compression and recovery should include an analysis of the thoracic as well as lumbar IVDs.
The exercise-induced reduction in IVD height and volume is likely to reflect increased hydrostatic pressure inside the disc and loss of fluid from the IVD. Because it is believed that the IVD is better able to distribute loading when it has greater hydration, these data suggest that running for 30 min reduces the IVDs potential to absorb spinal loads. Nevertheless, IVDs rely on diffusion and convection to exchange nutrients and metabolic by-products with surrounding vasculature (9); consequently, cyclical changes in intradiscal pressure are important to maintain IVD health. For example, running training has been demonstrated to increase extracellular matrix production and cell proliferation in the lumbar IVD of rats (7). Although physical activity is encouraged for spinal health, the optimum loading regime required to maintain the balance of matrix turnover in the IVD is currently unknown.
In conclusion, this study demonstrated, through repeated MRI examinations, that 30 min of moderate-intensity treadmill running reduces the height and volume of IVDs through the thoracic and lumbar regions of the spine. Furthermore, the relative losses in IVD height in the thoracic region were approximately twice the respective values for the lumbar region.
The authors wish to acknowledge Dr. Veena Olma and Dr. Derek D’Souza for the clinical supervision of the MRI procedures and Mr. Greg Capern (CQUniversity) for his technical assistance during the MRI analyses.
No external funding was received to support this research.
No conflicts of interest exist for any of the authors.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Ahrens SF. The effect of age on intervertebral disc compression during running. J Orthop Sports Phys Ther
. 1994; 20 (1): 17–21.
2. Beynon C, Burke J, Doran D, Nevill A. Effects of activity rest schedules on physiological strain and spinal load in hospital based porters. In: Reilly T, Greeves J, editors. Advances in Sport Leisure and Ergonomics
. London: Routledge; 2002. pp. 347–54.
3. Bonney RA, Corlett EN. Vibration and spinal lengthening in simulated vehicle driving. Appl Ergon
. 2003; 34 (2): 195–200.
4. Boocock MG, Garbutt G, Linge K, Reilly T, Troup JD. Changes in stature following drop jumping and post-exercise gravity inversion. Med Sci Sports Exerc
. 1990; 22 (3): 385–90.
5. Borg G. Perceived exertion as an indicator of somatic stress. Scand J Rehabil Med
. 1970; 2 (2): 92–8.
6. Bourne ND, Reilly T. Effect of a weightlifting belt on spinal shrinkage
. Br J Sports Med
. 1991; 55 (4): 209–12.
7. Brisby H, Wei AQ, Molloy T, Chung SA, Murrell GA, Diwan AD. The effect of running exercise on intervertebral disc extracellular matrix production in a rat model. Spine
. 2010; 35 (15): 1429–36.
8. Cavanagh PR, Lafortune MA. Ground reaction forces in distance running. J Biomech
. 1980; 13 (5): 397–406.
9. Chan S, Ferguson S, Gantenbein-Ritter B. The effects of dynamic loading
on the intervertebral disc. Eur Spine J
. 2011; 20 (11): 1796–812.
10. Corlett EN, Eklund JAE, Reilly T, Troup JDG. Assessment of workload from measurements of stature. Appl Ergon
. 1987; 18 (1): 65–71.
11. Dimitriadis AT, Papagelopoulos PJ, Smith FW, et al.. Intervertebral disc changes after 1 h of running: a study on athletes. J Int Med Res
. 2011; 39 (2): 569–79.
12. Dowzer CN, Reilly T, Cable NT. Effects of deep and shallow water running on spinal shrinkage
. Br J Sports Med
. 1998; 32 (1): 44–8.
13. Fowler NE, Lees A, Reilly T. Changes in stature following plyometric drop-jump and pendulum exercises. Ergon
. 1997; 40 (12): 1279–86.
14. Fryer JCJ, Quon JA, Smith FW. Magnetic resonance imaging and stadiometric assessment of the lumbar discs after sitting and chair-care decompression exercise: a pilot study. Spine J
. 2010; 10 (4): 297–305.
15. Garber CE, Blissmer B, Deschenes MR, et al.. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance for prescribing exercise. Med Sci Sports Exerc
. 2011; 43 (7): 1334–59.
16. Garbutt G, Boocock MG, Reilly T, Troup JD. Running speed and spinal shrinkage
in runners with and without low back pain. Med Sci Sports Exerc
. 1990; 22 (6): 769–72.
17. Han SM, Lee SY, Cho MH, Lee JK. Disc hydration measured by magnetic resonance imaging in relation to its compressive stiffness in rat models. Proc Inst Mech Eng H
. 2001; 215 (5): 497–501.
18. Kingsley M, D’Silva LA, Kilduff L. Direct evidence to demonstrate that running causes spinal shrinkage
within the lumbar vertebral region. J Sports Sci
. 2008; 26: S102–S103.
19. Leatt P, Reilly T, Troup JG. Spinal loading during circuit weight-training and running. Br J Sports Med
. 1986; 20 (3): 119–24.
20. Leivseth G, Drerup B. Spinal shrinkage
during work in a sitting posture compared to work in a standing posture. Clin Biomech
. 1997; 12 (7–8): 409–18.
21. Lewis SE, Fowler NE. Changes in intervertebral disk dimensions after a loading task and the relationship with stature change measurements. Arch Phys Med Rehabil
. 2009; 90 (10): 1795–9.
22. Marfell-Jones M, Olds T, Stewart A, Carter L. International Standards for Anthropometric Assessment
. Potchefstroom (South Africa): ISAK; 2006. p. 133.
23. Reilly T, Chana D. Spinal shrinkage
in fast bowling. Ergon
. 1994; 37 (1): 127–32.
24. Reilly T, Tyrrell A, TJ D. Circadian variation in human stature. Chronobiol Int
. 1984; 1 (2): 121–6.
25. Tyrrell AR, Reilly T, Troup JDG. Circadian variation in stature and the effects of spinal loading. Spine
. 1985; 10: 161–4.
26. van Deursen LL, van Deursen DL, Snijders CJ, Wilke HJ. Relationship between everyday activities and spinal shrinkage
. Clin Biomech
. 2005; 20 (5): 547–50.
27. Violas P, Estivalezes E, Briot J, Sales de Gauzy J, Swider P. Objective quantification of intervertebral disc volume properties using MRI in idiopathic scoliosis surgery. Magn Reson Imaging
. 2007; 25 (3): 386–91.
28. Violas P, Estivalezes E, Pedrono A, de Gauzy JS, Sevely A, Swider P. A method to investigate intervertebral disc morphology from MRI in early idiopathic scoliosis: a preliminary evaluation in a group of 14 patients. Magn Reson Imaging
. 2005; 23 (3): 475–9.
29. Yantzer BK, Freeman TB, Lee WE, et al.. Torsion-induced pressure distribution changes in human intervertebral discs. Spine
. 2007; 32 (8): 881–4.
30. Zhou SH, McCarthy ID, McGregor AH, Coombs RR, Hughes SP. Geometrical dimensions of the lower lumbar vertebrae–analysis of data from digitised CT images. Eur Spine J
. 2000; 9 (3): 242–8.
Keywords:©2012The American College of Sports Medicine
SPINAL SHRINKAGE; MECHANICAL LOADING; DYNAMIC LOADING; AXIAL COMPRESSION; HYDROSTATIC PRESSURE