The lumbar and sacrum movement pattern during the back squat exercise

THE LUMBAR AND SACRUM MOVEMENT PATTERN

DURING THE BACK SQUAT EXERCISE

MARK R. MCKEAN, PETER K. DUNN, AND BRENDAN J. BURKETT

Australian Institute of Fitness Research, School of Health and Sport Science, Faculty of Science, Health, and Education,University of Sunshine Coast, Queensland, Australia

ABSTRACT

McKean, MR, Dunn, PK, and Burkett, BJ. The lumbar and

sacrum movement pattern during the back squat exercise.

J Strength Cond Res 24(10): 2731–2741, 2010—An essential

exercise for strength training of the lower limbs is the squat

exercise. During this exercise, changes in lumbar lordosis are

commonly used to indicate when the descent of the squat

should cease, yet the behavior of the lumbar–scarum segments

remains unclear. The purpose of this study was to quantify the

lumbar–sacrum movements during the back squat, because the

movement of the sacrum is influenced by the width of stance,

this variable was also investigated. Thirty trained subjects, 18

men with 1 repetition maximum (1RM) squat of 123% (13.9%)

of bodyweight and 12 women with 1RM squat of 93% (15.6%),

performed a set of narrow and wide stance squats, each

carrying an additional 50% of body weight as load. The timing

and movement of the lumbar angle (T12/L1), sacrum angle (L5/

S1), and lumbar flexion angle (lumbar lordosis) were measured

in 3 dimensions for the ascent and decent phases. Men and

women achieved similar lumbar angles for both width of stance

and phase. Sacrum angles, lumbar flexion angles, and timing

differed significantly (p , 0.05) between gender and width of

stance. The lumbar flexion range during the descent phase for

women in narrow and wide stance was 12.9� and 12.6�,respectively; for men, this range was significantly (p , 0.05)

larger at 26.3� and 25.4�, respectively. Men and women

developed different movement patterns for the squatting

movement, and therefore, this needs to be considered in

strength development and screening procedures. The lumbar

spine became kyphotic as soon as a load was placed on the

shoulders, and any teaching cues to maintain a curved lumbar

spine when squatting must be questioned.

KEY WORDS lumbar, lumbar flexion, sacrum, segment timing

INTRODUCTION

One of the fundamental exercises for lower bodystrength, general fitness, and rehabilitation pro-grams is the squat. The majority of previousresearch on this exercise has been devoted to hip

and knee movements (16,32,35), possibly because of therelative ease of measuring the movements of these joints.Research of the lumbar spine and pelvic movements in squatshas been limited. Some research methodologies suggesta correct way to perform the squat (8,12), yet the techniquedescribed contains limited references to the actual move-ments of the lumbar and sacrum segments. The appropriatesquat technique is controversial, with suggestions that thelumbar curve should be maintained throughout the squat(9,20), whereas other research instructs the subjects tomaintain a ‘‘flat to arched but not rounded lumbar spine’’(28). In the few cases where subjects were instructed ona lumbar position when performing the squat, the lumbarcurve, or change in its position, was not actually monitored ormeasured during the performance. The majority of researchon squat technique provides no quantified measure or changein the position for the lumbar spine when squatting (13). Toensure the safety and effectiveness of this common exercise,initially quantifying the movement of the lumbar and sacrumregion is necessary.Most research into lumbar lordosis has been conducted in

the field of industrial squat lifting where the weight lifted is infront of the body (3,7,30). In this type of lifting, Burgess-Limerick recommends avoiding extreme lumbar vertebralflexion where flexion increases by .60� (7). For strengthtraining squats, McCaw and Melrose (26) and Liebenson (20)suggest the squat should be performed to full depth as long asthe lordotic curve is maintained. In all cases, the recommendedposture of the lumbar curve was not substantiated by anyquantification of the lumbar spine movements during theexercise. The alignment of the pelvis has also been found toinfluence the function of squat lifting, with an anterior tilt of thepelvis providing increased trunk muscle activity and thereforemore muscular support, or �core stability,� when squat liftingand lowering (10). However, the squat lifting technique is anindustrial type of lift, which allows the heels to lift off the floor,and the load is held in front of the body, which is quite differentfrom a strength training deep back squat technique.

Address correspondence to Mark R. McKean, [email protected].

24(10)/2731–2741

Journal of Strength and Conditioning Research� 2010 National Strength and Conditioning Association

VOLUME 24 | NUMBER 10 | OCTOBER 2010 | 2731

Copyright © National Strength and Conditioning Association Unauthorized reproduction of this article is prohibited.

Width of stance variations have long been used by thestrength coaches and fitness trainers (13,26) as a means toprovide a variation of the squat exercise. However, altering thewidth of stance has produced conflicting results with somestudies showing no noticeable change in muscle recruitment(13,31,33) and others finding width of stance did affect musclerecruitment patterns (26), whereas Escamilla et al. (13) showedno significant differences in trunk lean between 3 differentwidths of stance. In all these studies, there was no evidence ordiscussion regarding the influence of altered width of stance onthe lumbar spine and sacrum movements.In previous squat research studies, the loads have ranged from

body weight squats (1,11,32) to 1 repetition maximum (1RM)(26). Submaximal loadsminimize fatigue (8), provide a repetitionrange similar to those common in training (14), and reduce thelikelihood of changes to technique attributable to heavy loads,which allows a consistent performance across repetitions (31).Loads up to 1RM, or equal to 2.5 times a subject’s body weight,have been used in studies not specifically researching onmovement patterns and limb coordination.Gender differences in the pelvic dimensions (5,25) lumbar

vertebrae sizes (18), and trunk geometry (23) have beenreported in the literature. Research suggests that thesedifferences should be considered when developing bio-mechanical models for the lumbar spine (23), yet theinfluence of these gender differences on the movements ofthe lumbar and sacrum segments during the squat have notyet been identified.The concentric and eccentric contractions associated with

squatting break the movement into 2 very distinct phases, thedescent and the ascent.Walsh et al. (34) showed differences inlumbar behavior for each phase, and Escamilla et al. (13)found differences in forward trunk lean at similar knee flexionangles in each phase. The combined movement and criticalinteraction of the lumbar and sacrum region however has notyet been quantified, nor the influence of the descent andascent phases considered. Based on the previously estab-lished differences in concentric and eccentric movement,gender differences in pelvic anatomy, it was hypothesizedthat gender, phase, and width of stance would influence thelumbar–sacrum profile when performing the back squat.

Therefore, the specific aims of this study were to firstlyquantify in 3-dimensions, the timing and range of movementin the lumbar–sacrum region when performing the backsquat exercise and to secondly to determine the influence of(a) 2 different widths of stance (narrow and wide), (b) gender,and (c) the ascent and descent phases.

METHODS

Experimental Approach to the Problem

Thirty subjects completed 1 set of 8 back squat repetitions for2 widths of stance; (a) narrow stance in (NS) which the widthof stancewas equal to anterior superior iliac spine (ASIS)widthand (b) wide stance (WS) with stance twice ASIS width. Eachsquat was performed with an additional load equal to 50% ofthe subject’s body weight. The time and position of the lumbarspine and sacrum were measured and analyzed.

Subjects

All subjects freely volunteered to participate in the study.Subjects were informed of the experimental risks, and writteninformed consent was obtained under the guidelinesapproved by the University Human Research and EthicsCommittee before any experimental testing. The subjectswere either in their final year of study to become or alreadyworking as an accredited personal trainer. Subjects completeda questionnaire detailing training and medical history.Subjects were eliminated if they had not participated inregular strength training using the back squat exercises in thelast 12months or if they had any history ofmedical conditionsrelated to the back and lower extremities. Subjects had beenperforming squats in their strength and conditioning trainingprograms at least twice a week for a minimum of 12 months,and their relative mean (SD) squatting to body weight ratiowas 123% (13.9%) for men and 93% (15.6%) for women.Subject data are shown in Table 1.

Procedures

Standing height was measured to the nearest 1 mm usinga portable Stadiometer (Telescopic Metal Height Scale(PE063) Mentone International, Moorabin, Australia). Usingcalibrated electronic scales (TI BWB-600P—Digital PersonalScale, Wedderburn Pty Ltd, Willawong, Australia) total body

TABLE 1. Subject characteristics.

Subjects Age (y) Weight (kg) Height (cm) ASIS width (cm) 1RM as % of body weight

Women (n = 12) 24.2 (6.5) 62.1 (7.9) 167.1 (4.9) 24.4 (2.2) 93 (15.6)Men (n = 18) 24.1 (4.9) 83.2 (11.5) 179.5 (6.1) 25.2 (1.5) 123 (13.9)Combined (n = 30) 24.1 (5.5) 74.7 (14.6) 174.6 (8.3) 24.9 (1.8) 112 (17.3)

*1RM = 1 repetition maximum.†Values are given as mean (SD).ASIS = anterior superior iliac spine.

2732 Journal of Strength and Conditioning Researchthe TM

Lumbar and Sacrum Movement Pattern


mass was measured to the nearest 0.01 kg. Lumbar spine,sacrum, and lower limb motion data were collected in realtime at 120Hz by a 3 Dimensional Magnetic Tracking Device(Motion Monitor, Version 6.50.0.1 Innovative Sports Train-ing, Chicago, IL, USA). Eight magnetic sensors were placed

TABLE 2. Position of sensors for 3D magnetictracking device.

Sensor Anatomical segment

1 Stylus to digitize thesubjects dimensions

2 Thorax (the C7–T12 junction)3 Lumbar (dorsal surface

of T12–L1 junction)4 Sacrum (dorsal surface

of L5–S1 junction)5 Left thigh (anteromedial aspect

of upper thigh)6 Right thigh (anteromedial aspect

of upper thigh)7 Left shank (anteromedial aspect

of the tibia shaft)8 Right shank (anteromedial aspect

of the tibia shaft)

Figure 1. Digital picture of squat technique with sensors attached.

Figure 2. Digital picture of the sensor setup for the lumbar and sacrumspine.

TABLE 3. The ICC values for the dependentvariables.*

Descentphase

Ascentphase

NS WS NS WS

Men (n = 18)Lumbar angle 0.95 0.98 0.95 0.96Lumbar time 0.82 0.94 0.91 0.87Sacrum angle 0.96 0.94 0.95 0.97Sacrum time 0.90 0.82 0.93 0.92Lumbar flexion angle 0.93 0.94 0.92 0.95Lumbar flexion time 0.91 0.90 0.83 0.87

Women (n = 12)Lumbar angle 0.94 0.93 0.97 0.93Lumbar time 0.79 0.80 0.76 0.73Sacrum angle 0.92 0.95 0.95 0.94Sacrum time 0.81 0.87 0.74 0.84Lumbar flexion angle 0.96 0.98 0.98 0.99Lumbar flexion time 0.87 0.90 0.85 0.95

*ICC = intraclass correlation coefficient; NS = narrowstance; WS = wide stance.


Journal of Strength and Conditioning Researchthe TM

| www.nsca-jscr.org


TABLE 4. The start angles for lumbar, sacrum, and lumbar flexion in unloaded and loaded set ups for NS and WS squats.*†

Angles before commencement of the descent phase

NS unloaded NS + 50% BW WS unloaded WS + 50% BW

Men (n = 18)Lumbar start angle 70.8 (5.2) 71.0 (6.4) 75.6 (5.7) 70.8 (6.4)Sacrum start angle 43.7 (8.0) 46.6 (9.4) 47.0 (8.1) 43.1 (9.2)Lumbar flexion start angle 2.7 (11.7) 8.5 (11.4) 22.0 (10.3) 7.1 (12.5)

Women (n = 12)Lumbar start angle 79.8 (9.6) 69.2 (11.0) 72.8 (7.5) 69.8 (9.8)Sacrum start angle 42.8 (7.9) 47.2 (10.9) 48.3 (8.3) 46.7 (10.5)Lumbar flexion start angle 3.7 (7.9) 7.9 (4.8) 1.3 (7.1) 5.9 (5.2)

*BW = body weight; NS = narrow stance; WS = wide stance.†Values are given as mean (SDs) in degrees.

TABLE 5. The maximum angle for the lumbar and normalized time when the maximum occurred during the descent andascent phase of the back squat, for NS and WS squats with 95% CI size effect.*†

Descent phase Ascent phase

Description NS WS NS WS

Men (n = 18)Lumbar start angle (�) 71.0 (0.9) 70.8 (0.7) 37.7 (0.8)‡ 40.4 (1.2)‡95% CI (69.2, 72.7) (69.3, 72.3) (36.0, 39.4) (38.0, 42.9)Lumbar max angle (�) 36.8 (0.9)§ 40.0 (1.0)§ 37.3 (0.8)§ 40.0 (1.0)§95% CI (35.0, 38.5) (38.0, 42.0) (35.7, 38.9) (38.1, 42.0)Time max lumbar (%) 90.3 (1.2) 91.1 (1.2) 93.5 (1.9)k 89.1 (1.6)k

95% CI (87.9, 92.7) (88.6, 93.6) (89.7, 97.2) (85.9, 92.3)Lumbar range (�) 234.2 (1.2){ 230.8 (1.2){ 20.4 (0.3) 20.4 (0.6)95% CI (236.6, 231.8) (233.2, 228.4) (20.9, 0.1) (21.5, 0.8)

Women (n = 12)Lumbar start angle (�) 69.2 (1.8) 69.8 (1.6) 35.8 (1.1)# 39.7 (1.1)#95% CI (65.5, 72.9) (66.5, 73.1) (33.6, 38.0) (37.4, 41.9)Lumbar max angle (�) 35.5 (1.0)** 38.8 (1.1)** 33.8 (1.3)** 38.3 (1.0)**95% CI (33.4, 37.6) (36.6, 41.0) (31.1, 36.5) (36.3, 40.4)Time max lumbar (%) 94.1 (1.1) 92.8 (1.5) 93.6 (1.7) 91.7 (1.6)95% CI (91.9, 96.4) (89.7, 95.8) (90.1, 97.2) (88.5, 94.9)Lumbar range (�) 233.8 (1.4)†† 231.0 (0.9)†† 22.0 (1.0) 21.3 (0.5)95% CI (236.6, 230.9) (232.9, 229.1) (24.1, 0.1) (22.4, 20.3)

*CI = confidence interval; NS = narrow stance; WS = wide stance.†Values are given as mean (SE).‡Significant difference (p , 0.01) in the starting lumbar angle for men in the ascent phase, when comparing the width of stance.§Significant difference (p, 0.001) in the maximum lumbar angle achieved for men in both phases, when comparing the width of stance.kSignificant difference (p, 0.05) in the timing of maximum lumbar angle achieved by men in the ascent phase, when comparing the

width of stance.{Significant difference (p, 0.001) in the lumbar angle range achieved by men in the descent phase, when comparing the width of stance.#Significant difference (p , 0.001) in the starting lumbar angle for women in the ascent phase, when comparing the width of stance.**Significant difference (p, 0.001) in the maximum lumbar angle achieved for women in both phases, when comparing the width of

stance.††Significant difference (p, 0.001) in the lumbar angle range achieved by women in the descent phase, when comparing the width

of stance.




on anatomical segments as described in Table 2. Validation ofthe system was confirmed against standardized referencemeasures, and the variation was ,0.5� and within 0.0034 m.Three-dimensional magnetic tracking has been previouslyvalidated (27).The subjects were given time to complete their usual warm-

up procedure. With the sensors attached, the subjectscompleted 1 further warm-up set of squats with no additionalload. The width of each subject’s pelvis was measured betweenright and left ASIS, using skeletal goniometers (TTM BoneCaliper [PE054] Mentone International, Moorabin, Australia).Based on research by Escamilla et al. (14), narrow stance wasdefined when the inside distance between the subject’s heelsequaled the pelvic width, measured from right ASIS to leftASIS.Wide stance squat was defined as twice the pelvic width,measured from right to left ASIS. The width of each stancewas established using a steel measuring ruler and marks placedon the floor for each individual stance setup.

To determine if stance influenced movement patterns ina safe and repeatable manner, a submaximal load of bodyweight + 50% (BW+50%)was chosen. This is a common loademployed by beginners and women and allowed multiple setsof 8 repetitions without fatigue (12). Submaximal loads alsoreduce the possibility of nonvoluntary technique changesattributed to heavy loads (31) and allowed comparison toprevious studies. The capture of 3 blind repetitions for analysisoccurred within the set of 8 repetitions, the first and lastrepetitions were excluded (31).Foot alignment was not controlled, but in most cases in

the narrow stance squat, the feet were aligned parallelto the sagittal plane, that is, toes pointed straight ahead; and inthewide stance squat, the feet were between 20� and 30� awayfrom the midline. The 50% load was achieved by using analuminumOlympic bar placed across the upper trapezius andthe spine of scapula, with the subjects placing their hands ina palm forward grip.

TABLE 6. The maximum angle for the sacrum and normalized time when the maximum occurred during the descent andascent phase of the back squat, for NS and WS squats with 95% CI size effect.*†

Description


NS WS NS WS

Men (n = 18)Sacrum start angle (�) 46.6 (1.3)‡ 43.1 (1.2)‡ 33.2 (1.5)§ 32.9 (1.6)95% CI (44.1, 49.2) (40.7, 45.5) (30.3, 36.2) (29.7, 36.2)Sacrum max angle (�) 28.4 (1.3)k 28.7 (1.3) 30.8 (1.3)k 29.1 (1.4)95% CI (25.8, 31.0) (26.0, 31.3) (28.1, 33.4) (26.3, 31.9)Time max sacrum (%) 78.2 (1.9) 76.5 (1.9) 80.8 (2.7) 78.6 (2.3)95% CI (74.4, 81.9) (72.7, 80.2) (75.5, 86.2) (74.0, 83.1)Sacrum range (�) 218.2 (1.2){# 214.5 (1.0){# 22.4 (0.4)# 23.9 (0.9)#95% CI (220.7, 215.7) (216.4, 212.5) (23.3, 21.6) (25.6, 22.1)

Women (n = 12)Sacrum start angle (�) 47.2 (1.8) 46.7 (1.8) 23.1 (2.4)§ 27.1 (2.1)95% CI (43.5, 50.8) (43.1, 50.3) (18.2, 28.0) (22.9, 31.3)Sacrum max angle (�) 20.4 (1.9)k** 23.7 (1.3)** 20.3 (1.8)k††** 23.1 (1.3)††**95% CI (16.6, 24.2) (21.0, 26.3) (16.7, 24.0) (20.5, 25.6)Time max sacrum (%) 87.2 (1.9) 84.9 (1.7) 87.2 (2.3)‡‡ 82.5 (2.2)‡‡95% CI (83.4, 90.9) (81.5, 88.4) (82.6, 91.8) (78.0, 87.0)Sacrum range (�) 226.8 (1.1){§§ 223.1 (1.0){§§ 22.8 (0.9)§§ 24.0 (1.1)§§95% CI (228.9, 224.6) (225.1, 221.0) (217.9, 211.6) (26.3, 21.8)

*CI = confidence interval; NS = narrow stance; WS = wide stance.†Values are given as mean (SE).‡Significant difference (p, 0.001) in the starting sacrum angle for men in the descent phase, when comparing the width of stance.§Significant difference (p , 0.05) in the starting sacrum angle for NS squats in the ascent phase, when comparing gender.kSignificant difference (p , 0.05) in the maximum sacrum angle achieved in NS squats in both phases, when comparing gender.{Significant difference (p , 0.01) in the sacrum angle range for both WS and NS squats in the descent phase, when comparing

gender.#Significant difference (p , 0.05) in the sacrum angle range achieved for men in both phases, when comparing width of stance.**Significant difference (p, 0.001) in the maximum sacrum angle for women in both phases, when comparing the width of stance.††Significant difference (p , 0.001) in the starting sacrum angle for women in the ascent phase, when comparing the width of

stance.‡‡Significant difference (p , 0.01) in the timing of maximum sacrum angle achieved by women in the ascent phase, when

comparing the width of stance.§§Significant difference (p , 0.001) in the sacrum angle range achieved for women in both phases, when comparing width of

stance.



| www.nsca-jscr.org


Subjects assumed the squat start position unloadedwhile aninitial reading was taken. Then, with the loaded bar andcorrect width of stance, descended to the deepest point theyfelt comfortable with and in control. Subjects were allowed tostop squatting at any time they wished, and spotters wereready to assist the subject should the need arise. The actualtechnique or depth of the squat were not limited or controlled,but subjects were encouraged to perform the squat in thesame fashion they would do in normal training, and allsubjects squatted to a depth where thighs level with belowparallel. The squat was performed according to the NationalStrength and Conditioning Association (NSCA) guidelineson squats and monitored by the main researcher who is an

NSCA certified Certified Strength and Condiitoning Special-ist (CSCS) coach at the elite level. Three repetitions werecollected for analysis, and subjects were blind to which 3 ofthe 8 performed were recorded, although the first and lastrepetitions were excluded. Subjects were given 120-secondrecovery between sets. A digital image of the squat techniqueis shown in Figure 1.The magnetic tracking device measures the orthopedic

axes of the lumbar and sacrum. This instrument uses a singletransmitter strapped to the sacrum over L5/S1 and anothertransmitter strapped to the lumbar–thoracic junction over thespinous processes of T12/L1 as shown in Figure 2. These 2sensors feed back information regarding their relative

TABLE 7. The maximum angle for lumbar flexion and normalized time when the maximum occurred (mean and standarderror) during the descent and ascent phase of the back squat, for NS and WS with 95% confidence interval size effect.*

Description


NS WS NS WS

Men (n = 18)Lumbar flexion start (�) 8.5 (1.6)†‡ 7.1 (1.7)‡ 34.6 (2.9)†‡ 32.3 (2.8)‡95% CI (5.3, 11.6) (3.6, 10.5) (28.7, 40.4) (26.6, 38.0)Lumbar flexion max (�) 34.8 (2.9)§k 32.5 (2.8)k 34.9 (2.9)§k 32.7 (2.9)k

95% CI (28.9, 40.6) (26.9, 38.2) (29.1, 40.6) (26.9, 38.4)Time max lumbar flexion (%) 96.5 (1.9){ 97.9 (0.9) 97.2 (0.6){# 95.2 (0.9)#**95% CI (92.7, 99.3) (96.1, 99.6) (96.0, 98.3) (93.5, 97.0)Lumbar flexion range (�) 26.3 (1.8)†† 25.5 (1.7)†† 0.3 (0.1) 0.3 (0.1)95% CI (22.7, 29.9) (22.0, 28.9) (0.0, 0.6) (0.1, 0.5)

Women (n = 12)Lumbar flexion start (�) 7.9 (0.8)†‡‡ 5.9 (0.9)‡‡ 19.0 (1.9)† 18.3 (1.8)95% CI (6.3, 9.5) (4.2, 7.7) (15.1, 23.0) (14.7, 22.0)Lumbar flexion max (�) 20.8 (1.7)§§§ 18.5 (1.8)§§ 20.4 (1.9)§ 19.0 (1.8)95% CI (17.2, 24.3) (14.9, 22.2) (16.6, 24.2) (15.3, 22.7)Time max lumbar flexion (%) 83.6 (3.7){ 95.2 (2.0) 86.1 (3.0){ 88.3 (2.6)**95% CI (76.0, 91.1) (91.1, 99.3) (80.0, 92.2) (83.1, 93.6)Lumbar flexion range (�) 12.9 (1.8)††kk 12.6 (1.7)††kk 1.4 (0.5) 0.7 (0.2)95% CI (9.2, 16.5) (9.1, 16.1) (0.3, 2.5) (0.2, 1.1)

*NS = narrow stance; WS = wide stance.†Significant difference (p , 0.05) in the starting lumbar flexion angle for NS and WS squats in the ascent phase, when comparing

gender.‡Significant difference (p , 0.01) in the starting lumbar flexion angle for men in both phases, when comparing width of stance.§Significant difference (p , 0.05) in the maximum lumbar flexion angle achieved for NS squats in both phases, when comparing

gender.kSignificant difference (p, 0.01) in the maximum lumbar flexion angle achieved by men in the both phases, when comparing width

of stance.{Significant difference (p , 0.05) in the timing of maximum lumbar flexion angle in NS squat in both phases, when comparing

gender.#Significant difference (p , 0.01) in the timing of maximum lumbar flexion angle achieved by men in the ascent phases, when

comparing width of stance.**Significant difference (p, 0.05) in the timing of maximum lumbar flexion angle in WS squat in the ascent phase, when comparing

gender.††Significant difference (p , 0.01) in lumbar flexion range in NS and WS squat in the descent phase, when comparing gender.‡‡Significant difference (p, 0.001) in the starting lumbar flexion angle for women in the descent phase, when comparing width of

stance.§§Significant difference (p, 0.05) in the maximum lumbar flexion angle achieved by women in the descent phase, when comparing

width of stance.kkSignificant difference (p , 0.01) in the timing of maximum lumbar flexion angle achieved by women in the descent phase, when

comparing width of stance.




position to the world axis (lumbar and sacrum angles), andrelative to each other (lumbar flexion). Lumbar flexion wasmeasured comparing the change in relative angle betweenthese 2 sensors during the squat. An initial positivemeasurement would indicate a kyphotic lumbar curve andan initial negative measurement would indicate a lordoticlumbar curve. A reduction in this angle reflects a move

toward a more lordotic or increased curve in the lumbarspine and an increase in the lumbar flexion angle wouldindicate a move toward a flat or more kyphotic curve of thelumbar spine. The results reported show the absolute valuesfor the starting angle and maximum angle of each variableand the timing of when those angles were achieved.The time for each phase (descent and ascent) was

normalized, with the starting point at the top of the squat(highest vertical displacement of the sacrum) represented as0%, whereas the deepest part of the squat (lowest verticaldisplacement of the sacrum) represented as 100%, regardlessof the phase of the movement.

Statistical Analyses

A total of 6 different responses were analyzed (maximumlumbar angle; the time at which the maximum lumbar angleoccurred; maximum sacrum angle; the time at which themaximum sacrum angle occurred; the maximum lumbarflexion angle; and the time at which the maximum lumbarflexion angle occurred). Each of these 6 responses wereanalyzed separately for differences between gender (men orwomen), phase (ascent or decent), and stance (WS or NS).The results (Tables 4–7) are presented as mean and standarderror analyzed for gender (men or women), stance (WS orNS), and phase (ascent or decent). Intraclass correlationcoefficients (ICCs) were calculated to assess the reliability of

TABLE 8. Lumbar–pelvic ratios in trunk movement fordescent and ascent phases in NS and WS squatsby gender.*†

Descentphase

Ascentphase

NS WS NS WS

Men (n = 18) 1.88 2.12 0.17 0.10Women (n = 12) 1.26 1.34 0.71 0.32

*NS = narrow stance; WS = wide stance.†The ratio is calculated by dividing the lumbar range of

movement by the sacrum range of movement to showrelative movement of each segment towards total trunkinclination.

Figure 3. Sample male lumbar angle, sacrum angle, and lumbar flexion angle movement in an NS squat.



| www.nsca-jscr.org


the repeated measures using the Bland and Altman method(4). Intraclass correlation coefficient values ,0.4 representedpoor reliability, 0.4–0.7 fair, 0.70–0.90 good, and .0.9excellent reliability (15). One hundred percent of ICC valueswere in the good category (.0.7) and 68.8% in the excellentcategory (.0.9), showing good reliability of the data.Statistical interpretation focused on the main effects andthe threshold for statistical significance was set to p # 0.05.Linear mixed models were used to model the data usingsubjects as random effects. Models were fitted individuallywith 1 response and 1 explanatory variable (gender, phase, orstance) as a fixed effect. Differences were then detected usingthe p values from the analysis of variance F-tests on thecorresponding fixed effect.The ICC for the dependent variables is included in Table 3.

RESULTS

The initial angles for the lumbar, sacrum, and lumbar flexionangles for unloaded and loaded setups are presented in Table 4.These results show the adjustment made by the subjects to

the lumbar, sacrum, and lumbar flexion angles in response tothe addition of the loaded bar. Men adjusted lumbar andsacrum angles more in the wide stance squat in response to theload, whereaswomen adjusted these anglesmore in the narrowstance squat. For unloaded and loaded, wide and narrow stancesetup positions, men and women adjusted lumbar flexionangles so as to create a more kyphotic lumbar spine as a meansof coping with the additional load before squatting.Themaximum lumbar angle and normalized timewhen the

maximum occurred during the descent, ascent for both menand women are presented in Table 5.The maximum sacrum angle and normalized time when

the maximum occurred during the descent, ascent for bothmen and women are presented in Table 6.The maximum lumbar flexion angle and normalized time

when the maximum occurred during the descent, ascent forboth men and women are presented in Table 7.Lumbar and sacrum movements and the timing of when

the maximums occur in these movements differ significantlyacross gender andwidth of stance. Lumbar starting angles andranges of movement were similar across gender for eachstance, but the maximum lumbar angle achieved was differentintragender with both groups achieving a significantly re-duced maximum angle and range of movement with the widestance squat compared to the narrow squat.Women achieveda greater range of movement of the sacrum angle in thedescent phase and the timing of this maximum also occurredsignificantly earlier than for men.There were significant differences in techniques when

comparing the effect of width of stance. The wider stanceallowed all subjects to achieve reduced lumbar, sacrum, andlumbar flexion angles, whereas the narrow stance squatcaused increased lumbar and sacrum angles and increasedlumbar flexion.Men reachedmaximum sacrum angles soonerin the descent and later in the ascent for both widths of stance.

In lumbar flexion timing, women reached the maximumposition sooner in descent and later in ascent for both widthsof stance. The lumbar–sacrum ratios showing contributiontoward total trunk inclination in the squat are presented inTable 8.Men showed a higher lumbar–sacrum ratio for both widths

of stance. Men also increased the lumbar–sacrum ratiosubstantially for the wide squat compared to the narrowsquat. Women achieved similar ratios regardless of width ofstance. A sample of the typical movements of lumbar angle,sacrum angle, and lumbar flexion for both phases of a narrowsquat for a man is shown in Figure 3.

DISCUSSION

The first aim of this study was to quantify lumbar and sacrummovements when performing the back squat exercise andsubsequently determine if width of stance influenced thesemovements with regard to gender and phase.The unloaded spine positions of the sacrum were similar

across gender and stance width. However, the lumbar startingangles differed significantly between gender for both widths ofstance and lumbar flexion angles differed between genders inthe wide stance unloaded start position. Regardless of stanceand starting position, the adjustment of the lumbar, sacrum andlumbar flexion angles from the unloaded to loaded startpositions suggest that the human body has a means ofdetermining the ideal posture of the spine, thus allowing it toachieve the optimal position to manage the additional load.This premise is supported by Martin and Nelson (24) whosuggests that the trunk moves forward to align the combinedsubject and bar load vertically over the center of gravitycausing the spinal curve to decrease the lordosis.Once loaded with the 50% additional weight, men and

women had similar lumbar starting angles before the descentphase, with a difference of ,1.8� for the narrow and widestance squats. Sacrum starting angles were also similarbetween genders in the narrow squat, but men significantlyaltered their starting sacrum angle for wide stance squats,whereas women maintained similar sacrum angles regardlessof width of stance. Men appear to alter their sacrum angle toaccommodate either structural or mechanical differences inthe pelvic girdle. This significantly different sacrum startingangle between the 2 widths of stance for men results in amoreposterior tilt to the sacrum for narrow squats as shown by theincreased sacrum starting angle. Width of stance also hada significant difference in the starting lumbar flexion angle forboth genders, with a more kyphotic lumbar curve in thenarrow squat setup position than for the start of the widesquat position. This feature may be explained by the reportedincrease size of lumbar vertebral bodies and distance betweenlumbar vertebrae (18), and the narrower taller pelvis found inmen (25) which may cause men to stand more uprightthrough the trunk than women.In performing the squats, there were significant differences

in the maximum lumbar angle and resultant lumbar range of




movement. Both men and women achieved a significantlysmaller lumbar angle in the wide stance squat compared tothe narrow squat in both phases. Although these angles arevery small, this significant difference suggests that the widesquat allows the lumbar position to remain slightly moreupright and not lean as far forwards as with the narrow squat.In addition, there is also a significant difference for gender inthe range of lumbar motion between the 2 squats stances inthe descent phase, showing that men perform the narrowstance by allowing more forward lean of the lumbar spinethan in the wide stance squat. The difference in timing for themaximum lumbar angle reached by men in the ascent phaseof the narrow squat may also be because of the genderdifferences in lumbar structure mentioned previously.Women subjects reached themaximum forward position at

a later stage of the squat descent than men and significantlylater in the ascent for wide stance squats, suggesting the widerstance provides women with a greater range of flexibility inthe deepest part of the squat. Furthermore, this interestingresult indicates sacrum movement timing in the ascent canoccur much later, creating 2 different pathways for descentand ascent in wide stance squats. This feature may beattributed to the proportionally wider more oval pelvis andshorter lumbar vertebral length of women (18,25) and theaccepted increased level of flexibility of women (2,17).Lumbar flexion ranges have been shown to change with

torso inclination (22), and the greater the trunk inclination,the more likely the lumbar curvature will become kyphotic.Maximum lumbar flexion range of movement of 60� has beenreported (29), whereas in the current study for both widths ofstance, the start position for the lumbar spine was already ina kyphotic rather than lordotic position, with the lumbarflexion angle ranging from 5.9� to 8.5�. During the squatdescent, this kyphosis increased for both genders and in bothwidths of stance. This is supported by Walsh et al. (34), whofound that ‘‘weightlifting using a squat bar causes athletes tosignificantly hyperextend their lumbar spines.’’ In the currentstudy, men significantly increased their lumbar flexion andnearly doubled the lumbar flexion range of movementachieved by women, a finding that is also supported byWalshet al. To squat, it appears that men have a limited range ofmovement at the sacrum, and this is compensated by anincreased range of movement in lumbar flexion. Thisdecreased movement of the sacrum requires men to achievean increased kyphosis as part of their movement pattern. Thismay relate to the increased level of inherent stiffness menhave in the lumbar spine, compared to women (6). Womenappear to maintain a greater stiffness in the lumbar flexionangle and make up for this by increased range of movementin the sacrum. This lower level of inherent stiffness in thelumbar spine may necessitate women to establish moremuscular control and stability of the lumbar spine and thuskeep a stiffer lumbar curve in movements such as squatting.The timing of the different spine segments demonstrates

the different coordination methods between men and

women. In the descent phase, for both stance widths, menreach maximum angles of the 3 measures in the order, sacrumangle, lumbar angle, and lumbar flexion angle. In the ascentphase, men simply reversed the order. For women in thenarrow stance squat, the sequence in the descent phase waslumbar flexion angle, sacrum angle, and lumbar angle. Witha different pattern in the wide stance squat of sacrum angle,lumbar angle, and lumbar flexion angle. A different patternemerged in the narrow stance ascent phase with the orderlumbar angle, sacrum angle, and lumbar flexion angle. Finally,there was a different pattern again in the wide stance, with theorder of lumbar angle, lumbar flexion angle, and sacrum angle.The reason for this variation found among the women isunknown and is a topic for future studies.Lumbar–sacrum ratios were also calculated. This interac-

tion of the lumbar and sacrum described by the lumbar–sacrum ratio shows the differences between genders for themanner in which they achieve trunk inclination duringsquatting. Men show a significant difference in lumbar–sacrum ratios with more lumbar movement in the descentphase and less lumbar movement in the ascent than dowomen. Women achieved a more similar lumbar–sacrumratio across width of stance in the descent, whereas mendecreased sacrum movement in the wide stance squat, thusincreasing the lumbar–sacrum ratio. This is supported byother research studies (19,21,30) that show differences in thepatterns of movement and coordination between genders. Asdemonstrated in the current study, the segmental coordina-tion within the squat exercise is a complex issue, pre-dominately as there are many joints, levers, and structuralaspects that contribute to the overall movement patterns.Because of these many aspects and the differences inphysique, men and women differ in the manner in whichthey coordinate the lumbar sacral region to perform a squat.

PRACTICAL APPLICATIONS

Previous research has not reported on squat movementpattern differences between men and women. The currentresearch has shown that there are significant differences in thesquatting behavior for men and women when comparedacross both WS and NS setups. In the squat exerciseenvironment, professionals need to understand the influencethat the width of stance plays on these behaviors and thedifferent movement patterns men and women achieve underloads equal to 50% of BW. This would be most evident whenviewing women in the sagittal plane where health profes-sionals would see differences in the timing of the sequencesbetween descent and ascent phases when considering themovements of the sacrum and lumbar spines. Further, the useof the same screening protocols for squatting movementsbetween men and women must be questioned. The squat iscommonly used as a screening tool in the practical setting andcoaches and trainers should avoid comparing male squatpatterns to that of women especially when consideringscarum and lumbar movements and timing.



| www.nsca-jscr.org


The common squat technique presented in previousresearch suggests maintaining a flat to semicurved lumbarspine when squatting. The current research has shown that assoon as the 50% BW load is placed across the shoulders, thelumbar curve flattens and becomes slightly kyphotic beforethe subject commences the descent. This change of lumbarcurve is marginal, and we suggest it would be difficult to viewwith the naked eye. Coaches using the change in lumbarcurve, or lumbar flexion angle, as a determining factor, orteaching cue for good squat technique may in fact beinterfering with normal lumbar movement behavior. Further,the point at which the subject loses the lumbar curve cannotbe used as a cue to determine when a person should cease thedescent.We suggest that kyphosis of the lumbar spine in deepsquatting is a natural part of the squat movement when usingloads equal to 50% BW and coaches should not preventexperienced squatters from allowing this to happen to thesmall extent shown in this research.The current research also shows that the movement

pattern used by women when squatting involves a muchstiffer lumbar spine position or less change in the lumbarflexion. Previous research suggests men have a stiffer lumbarspine because of the additional soft tissue support thanwomen, and the authors suggest that women may benefitfrom enhanced strength training of the surrounding muscu-lature of the lumbar spine and trunk region to provideadditional stiffness to support the lumbar spine region duringsquat type activities. Coaches may see improved squattingabilities in women who develop supporting muscles of thetrunk during the preparation phases or before each pro-gressive increase in training loads.Finally width of stance created several changes in squatting

movement patterns across both men and women. Modifyingstance width is a common exercise variation in the strengthtraining environment, and this research has shown that themovement patterns so alter significantly with regards to thedifferent spinal segments. Coaches need to appreciate thiswhen altering stance width and not expect a person toperform the squat in an identical pattern of movement. Thisalso suggests that themotor pattern developed by peoplewhosquat is different between stances and both widths of stanceshould be taught progressively to ensure the movementpattern is equally learned across both widths of stance.

ACKNOWLEDGMENT

No funding was received for this research project.

REFERENCES

1. Abelbeck, KG. Biomechanical model and evaluation of a linearmotion squat type exercise. J Strength Cond Res 16: 516–524, 2002.

2. Allander, E, Bjornsson, O, Olafsson, O, Sigfusson, N, andThorsteinsson, J. Normal range of joint movements in shoulder, hip,wrist and thumb with special reference to side: A comparisonbetween two populations. Int J Epidemiol 3: 253–261, 1974.

3. Bazrgari, B, Shirazi-Adl, A, and Arjmand, N. Analysis of squat andstoop dynamic liftings: Muscle forces and internal spinal loads. EurSpine J 16: 687-699, 2006.

4. Bland, M and Altman, D. Measurement error and correlationcoefficients. BMJ 313: 41–42, 1996.

5. Brinckmann, P, Hoefert, H, and Jongen, H. Sex differences in theskeletal geometry of the human pelvis and hip joint. J Biomech 14:427–430, 1981.

6. Brown, M, Holmes, D, Heiner, A, and Wehman, K. Intraoperativemeasurement of lumbar spine motion segment stiffness. Spine 27:954, 2002.

7. Burgess-Limerick, R. Squat, stoop, or something in between? In:Straker, L and Pollock, C, (eds.) 1999. CD-ROM Proceedings ofCybErg 1999: The Second International Cyberspace Conference onErgonomics. The International Ergonomics Association Press,Curtin University Perth, 1999. pp. 143–148.

8. Caterisano, A, Moss, RF, Pellinger, TK, Woodruff, K, Lewis, VC,Booth, W, and Khadra, T. The effect of back squat depth on the emgactivity of 4 superficial hip and thigh muscles. J Strength Cond Res 16:428–432, 2002.

9. Comfort, P and Kasim, P. Optimizing squat technique. Strength CondJ 29: 10–13, 2007.

10. Delitto, RS and Rose, SJ. An electromyographic analysis of twotechniques for squat lifting and lowering. Phys Ther 72 438, 1992.

11. Dionisio, VC, Almeida, GL, Duarte, M, and Hirata, RP. Kinematic,kinetic and EMG patterns during downward squatting.J Electromyogr Kinesiol 18: 134–143, 2008.

12. Donnelly, DV, Berg, WP, and Fiske, DM. The effect of the directionof gaze on the kinematics of the squat exercise. J Strength Cond Res20: 145–150, 2006.

13. Escamilla, RF, Fleisig, GS, Lowry, TM, Barrentine, SW, andAndrews, JR. A three-dimensional biomechanical analysis of thesquat during varying stance widths. Med Sci Sports Exerc 33: 984,2001.

14. Escamilla, RF, Fleisig, GS, Zheng, N, Lander, JE, Barrentine, SW,Andrews, JR, Bergemann, BW, and Moorman III, CT. Effects oftechnique variations on knee biomechanics during the squat and legpress. Med Sci Sports Exerc 33: 1552, 2001.

15. Fleiss, J. The Design and Analysis of Clinical Experiments. New York,NY: John Wiley and Sons, 1986.

16. Fry, AC, Smith, JC, and Schilling, BK. Effect of knee position on hipand knee torques during the barbell squat. J Strength Cond Res 17:629–633, 2003.

17. Gabbard, C and Tandy, R. Body composition and flexibility amongprepubescent males and females. J Hum Mov Stud 14: 153–159,1988.

18. Gilsanz, V, Boechat, M, Ramon Gilsanz, P, Loro, M, Roe, T, andGoodman, W. Gender differences in vertebral sizes in adults.Radiology 190: 678–682, 1994.

19. Kernozek, TW, Torry, MR, Van Hoof, H, Cowley, H, and Tanner, S.Gender differences in frontal and sagittal plane biomechanics duringdrop landings. Med Sci Sports Exerc 37: 1003, 2005.

20. Liebenson, C. Activity modification advice: Part II–squats. J BodyMov Ther 7: 230–232, 2003.

21. Lindbeck, L and Kjellberg, K. Gender differences in lifting technique.Ergonomics 44: 202–214, 2001.

22. Maduri, A, Pearson, B, and Wilson, S. Lumbar-pelvic range andcoordination during lifting tasks. J Electromyogr Kinesiol 18: 807–814,2007.

23. Marras, W, Jorgensen, M, Granata, K, and Wiand, B. Female andmale trunk geometry: Size and prediction of the spine loading trunkmuscles derived from MRI. Clin Biomech (Bristol, Avon) 16: 38–46,2001.




24. Martin, P and Nelson, R. The effect of carried loads on the walkingpatterns of men and women. Ergonomics 29: 1191–1202, 1986.

25. Mays, S and Cox, M. Sex Determination in Skeletal Remains. HumanOsteology in Archaeology and Forensic Science. New York: CambridgeUniversity Press, 2000. pp. 117–142.

26. McCaw, ST and Melrose, DR. Stance width and bar load effects onleg muscle activity during the parallel squat. Med Sci Sports Exerc 31428, 1999.

27. Mills, P, Morrison, S, Lloyd, D, and Barrett, R. Repeatability of 3Dgait kinematics obtained from an electromagnetic tracking systemduring treadmill locomotion. J Biomech 40: 1504–1511, 2007.

28. Morrissey, MC, Harman, EA, Frykman, PN, and Han, KH. Earlyphase differential effects of slow and fast barbell squat training. Am JSports Med 26: 221, 1998.

29. Norkin, C, White, D, and White, J. Measurement of Joint Motion: AGuide to Goniometry. Philadelphia, PA: FA Davis, 1995.

30. O’Brien, A and O’Sullivan, L. A biomechanical, physiological andpsychophysical study of the squat, stoop and semi-squat liftingtechniques. In: Proceedings of the Irish Ergonomics Society Annual

Conference. O’Sullivan, L and Chan, S, eds. Limerick, Ireland:University of Limerick, 2005. pp. 26–31.

31. Paoli, A, Marcolin, G, and Petrone, N. The effect of stance width onthe electromyographical activity of eight superficial thigh musclesduring back squat with different bar loads. J Strength Cond Res 23:246, 2009.

32. Scaglioni-Solano, P, Song, JE, and Salem, GJ. Lower extremitybiomechanics during different squat depths. Med Sci Sports Exerc 37:S393, 2005.

33. Signorile, JF, Kacsik, D, Perry, A, Robertson, B, Williams, R,Lowensteyn, I, Digel, S, Caruso, J, and LeBlanc, WG. The effect ofknee and foot position on the electromyographical activity of thesuperficial quadriceps. J Orthop Sports Phys Ther 22: 2–9, 1995.

34. Walsh, JC, Quinlan, JF, Stapleton, R, FitzPatrick, DP, andMcCormack, D.Three-dimensional motion analysis of the lumbar spine during ‘‘Freesquat’’ Weight lift training. Am J Sports Med 35: 927, 2007.

35. Wretenberg, PER, Feng, YI, and Arborelius, U. High-and low-barsquatting techniques during weight-training.Med Sci Sports Exerc 28:218, 1996.



| www.nsca-jscr.org


Documents

The lumbar and sacrum movement pattern during the back squat exercise