Human Movement Science 31 (2012) 897–906

Contents lists available at SciVerse ScienceDirect

Human Movement Science

journal homepage: www.elsevier .com/locate/humov

Effects of overweight and obesity on walkingcharacteristics in adolescents

Janet S. Dufek a,⇑, Rayland L. Currie a,1, Philana-Lee Gouws a,1, Lori Candela b,2,Antonio P. Gutierrez b,2, John A. Mercer a,1, LeAnn G. Putney c,3

a Biomechanics Laboratory, Department of Kinesiology and Nutrition Sciences, 4505 South Maryland Parkway, Box 3034,University of Nevada, Las Vegas, Las Vegas, NV 89154-3034, United Statesb Department of Psychosocial Nursing, 4505 South Maryland Parkway, Box 3018, University of Nevada, Las Vegas, Las Vegas,NV 89154-3018, United Statesc Department of Educational Psychology, 4505 South Maryland Parkway, Box 3303, University of Nevada, Las Vegas, Las Vegas,NV 89154-3303, United States

a r t i c l e i n f o

Article history:Available online 9 December 2011

PsychINFO classification:2330

Keywords:Body mass indexChildrenGaitLocomotionSpatio-temporal parameters

0167-9457/$ - see front matter � 2011 Elsevier B.doi:10.1016/j.humov.2011.10.003

⇑ Corresponding author. Tel.: +1 702 895 0702.E-mail address: janet.dufek@unlv.edu (J.S. Dufe

1 Tel.: +1 702 895 0702.2 Tel.: +1 702 895 2443.3 Tel.: +1 702 895 4879.

a b s t r a c t

Child and adolescent obesity is growing at a staggering rate. Associ-ated potential health risks have been acknowledged in the adultpopulation, and similar concerns have been raised for children andadolescents. However, distinguishing locomotor characteristicsrelated to obesity have yet to be clearly identified for adolescents.The aims of the study were to examine the effects of walking velocityand gender on spatio-temporal characteristics of gait between nor-mal weight (NW) and overweight and obese (OWO) adolescents. Inaddition, we sought to identify characteristics of gait that are relatedto body mass index percentile (BMI%). Adolescent students in grades7–10 (N = 111) from a charter school participated in the study. Allparticipants walked at two speeds (preferred, fast) over an instru-mented walkway (120 Hz). Spatio-temporal characteristics of gaitwere extracted from four walkway passes and evaluated. Resultsidentified significant (p < .05) differences in velocity, percent doublesupport, percent swing phase, and stance width between groups.Only stance width was different (p < .05) between genders. Modelsto predict BMI% from kinematic walking parameters were of moder-ate strength (averaging 43.5% explained variance) and were gener-ally stronger for females versus males. Percent double support wasthe primary predictor variable of BMI% across speed and gender. It

V. All rights reserved.

k).

898 J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906

is suggested that OWO adolescents may be challenged with controlof movement of the center of mass during the support phase ofwalking.

� 2011 Elsevier B.V. All rights reserved.

1. Introduction

Pediatric obesity is emerging as a 21st century global epidemic. It is currently estimated that 155million children worldwide, or 1 out of 10, are obese or overweight (Aschemeier, Lange, Kordonouri, &Danne, 2008). In the United States, childhood obesity has tripled in the last 20 years with 1 in 5 chil-dren now considered overweight (Daniels et al., 2005; Dehghan, Akhtar-Danesh, & Merchant, 2005;Whitaker, Davis, & Lauer, 2004). Ethnicity also appears to influence child obesity statistics with thegreatest number of overweight and obese children found disproportionately among Native American,African American, and Hispanic populations (Caprio et al., 2008). In addition, up to 70% of obese chil-dren mature to become obese adults (Dehghan et al., 2005). Furthermore, other obesity-related healthconcerns include type 2 diabetes, cardiovascular disease, and respiratory challenges (Lawrence &Kopelman, 2004), while the combination of adulthood and obesity has also been shown to be stronglyassociated with the onset of bilateral knee joint osteoarthritis (Anderson & Felson, 1988; Stürmer,Günther, & Brenner, 2000).

In adults, reported effects of obesity on gait characteristics and potential joint pathology have beenequivocal. Devita and Hortobágyi (2003) reported greater knee and ankle joint moments for obese par-ticipants walking at a standard speed (1.50 m/s) versus a preferred speed (1.29 m/s). Despite the dif-ferences in body mass and assumed increases in joint loads for obese individuals, hip and knee jointmoments were not different for obese versus non-obese adults at the matched walking speed. In con-trast, Browning and Kram (2007) identified peak extensor knee moments to be 51% greater in obeseversus non-obese adults walking at 1.50 m/s. Of note is the fact that as walking speed decreased,the magnitude of the joint moment differences between groups became less and were eventuallynon-existent, suggesting obese individuals can modulate joint loads by self-selecting walking speed.The contrasting joint kinetic profiles between groups were previously suggested by Devita andHortobágyi (2003) to be the result of altered kinematics observed for the obese subjects. This findingwas subsequently suggested to be a conscious effort to increase dynamic balance to reduce knee jointloads (Wearing, Hennig, Byrne, Steele, & Hills, 2006).

However, Browning and Kram (2007) identified no differences in kinematics; alternatively, theirresults showed increased ground reaction forces leading to increased joint kinetics for obese adults.Malatesta et al. (2009) extended this body of work to examine mechanical work optimization forobese adults walking at a preferred speed. The authors reported greater external mechanical workfor obese versus normal weight individuals, attributable to the mass differential. Furthermore, resultsdid not suggest that obese adults modify kinematics to improve energy transfer, but rather choose aslower walking speed to minimize mechanical work associated with walking.

It is not known if obese children display similar kinematic or kinetic profiles as obese adults, and assuch, locomotor characteristics of obese children are now receiving research attention. Hills andParker (1992) examined kinematic differences between obese (n = 10) and normal weight (n = 4) chil-dren and found lesser velocities, increased instability, and limb asymmetry for obese children acrosstheir small study sample. McMillan, Auman, Collier, and Williams (2009) reported that overweightboys (n = 7) displayed greater knee valgus and hip adduction during support than healthy weight boys(n = 7), and suggested such frontal plane excursions in the overweight individuals will lead to lowerextremity injuries and performance deviations. Paralleling the queries into walking speed reportedfor obese adults, Shultz, Sitler, Tierney, and Hillstrom (2009) reported no kinematic differences duringstance or swing between normal weight (n = 10) and overweight (n = 10) children with selected differ-ences identified across walking speed. Lower extremity joint moment values were greater in the over-weight group, with no differences in percent increase with increased walking speed. In contrast,

J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906 899

McMillan, Pulver, Collier, and Williams (2010) did find differences in velocity-matched healthy weight(n = 18) and obese (n = 18) adolescents, suggesting that obese participants adopted a kinematic strat-egy to minimize hip and knee joint moments.

The effects of gender on performance of locomotor skills in children have resulted in equivocalfindings. Haubenstricker, Wisner, Seefeldt, and Branta (1997) identified differences in running perfor-mance between boys and girls. In contrast, Barnett, Beurden, Morgan, Brooks, and Beard (2010) foundno gender differences in locomotor skills in adolescents. It has been suggested that locomotor skilldevelopment may be more directly related to other factors, including weight (Oakley, Booth, & Chey,2004). Peyrot et al. (2010) reported no gender differences in obese adolescents before a weight lossprogram for a small sample (n = 16), yet little data exists examining gender differences and adiposityrelative to potential differences in developed locomotor skills, including walking, for adolescents for alarger sample size. It is also unknown if there are gender differences relative to walking performancerelative to body composition for this age group.

While the potential health challenges related to obesity are recognized (Lawrence & Kopelman,2004), there are conflicting bodies of evidence being presented in the literature relative to obesityand locomotion. Such results are likely influenced by experimental protocols, measurement proce-dures, data reduction techniques, participant selection and study sample size. We sought to addresssome of these potential limitations and possible sources of conflicting information by identifying alarge sample of ethnically diverse children who would participate in the study by virtue of the aca-demic curriculum offered at their school versus selection or voluntary participation. We also soughtto explore potential relationships between gender and obesity relative to adolescent gait.

Therefore, the specific aims of the study were to: (1) evaluate spatio-temporal characteristics oflocomotion between normal weight (NW) and overweight and obese (OWO) adolescents walking atdifferent speeds, (2) evaluate spatio-temporal characteristics of locomotion between NW and OWOadolescents across gender walking at different speeds, and (3) predict body mass index percentile(BMI%) from spatio-temporal locomotion characteristics of adolescents. The significance of achievingthe study aims includes understanding the potential long-term effects of overweight/obesity for ado-lescents as they move into adulthood, with an eye toward long-term overuse injury prevention. Wehypothesized that adiposity and walking speed would influence spatio-temporal parameters of gait;however, there would be no gender-adiposity effects. We further hypothesized that spatio-temporalparameters of gait would strongly predict level of adiposity for adolescents.

2. Methods

2.1. Participants

The entire body of students (N = 154) enrolled in grades 7–10 inclusive at an ethnically diversecharter school participated in data collection as a part of the curriculum at the school. Of this group,institutionally approved parental permission and child assent were obtained from 111 participants,constituting the entire study sample. Participants were between the ages of 12 to 17 years(14.2 ± 1.4 yrs) and enrolled in grades 7 (15 male, 15 female), 8 (21 male, 15 female), 9 (13 male, 9female) and 10 (15 male, 8 female). The self-reported ethnicity of this study sample is presentedgraphically in Fig. 1.

2.2. Data acquisition

All children performed walking gait assessment as part of a special curricular program (instruc-tional modules focusing on a modification to the Diabetes Prevention Program). During a regularlyscheduled academic physical activity period, height recorded to the nearest 1/8 inch (0.32 cm) andweight recorded to the nearest 1/4 pound (1.1 N) of each participant was obtained in a private screen-ing room. A trained registered nurse conducted all anthropometric measurement sessions using a dig-ital medical beam scale (HealthOMeter, 500 KL).

Fig. 1. Ethnicity of study participants as self-reported. Note: Af-Am = African American, As-Am = Asian American, Cauc = Cau-casian, Hisp = Hispanic, DNR = did not report.

900 J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906

Each participant then walked twice at their preferred pace over a 4.27 m instrumented walkway(CIR Systems, Haverhill, PA; 120 Hz), which was centered in a 10 m hallway. Reliability of the walkwaysystem (GAITRite) has been shown to be high to excellent (ICC(1,1) = .73 to .95) when used to evaluatea clinical population (Sorsdahl, Moe-Nilssen, & Strand, 2008). The walkway had an imperceptiblethickness of 4 mm which created no perturbation in the case where one’s foot made only partial con-tact with the mat. All participants started at the same position, walked through the hallway (over thewalkway), stopped, turned around, and walked back to the start position. This was followed by twotrials over the walkway at a fast pace, after receiving the following verbal instruction: ‘‘Walk as fastas you can.’’ There was no attempt to control footwear, velocity, or attire that the children wore whilewalking in order to more closely replicate typical movement patterns. A minimum of one full walkingstride was completed before and after contact with the walkway to minimize the effects of initialacceleration and terminal deceleration.

2.3. Data reduction

Body mass index (BMI) was calculated for each participant based upon height, weight, age, andgender, following the procedures established by the Centers for Disease Control and Prevention(CDC & Prevention, 2009). Participants were then grouped (post-hoc) based upon their BMI percentile(BMI%), again following the recommendations of the CDC which has established <85th percentile asnot overweight or obese, P85th percentile as overweight, and P95th percentile as obese. All partic-ipants with BMI% <85th percentile were assigned to the NW group (n = 56) and those with BMI%P85th percentile were assigned to the OWO group (n = 55).

Custom walkway software (GAITRite, version 4.0, 59) was used to visually inspect each walkingtrial performed for completeness. If any part of initial or terminal foot contact was not complete(the entire foot did not contact the walkway), that step was eliminated from the walking trial. Thisprocess resulted in a variable number of steps for each participant-condition-trial that was influencedby foot placement on the walkway and walking velocity. The range of steps was between 4 and 7across all trials and participants. The two walking trials at each speed were then concatenated, result-ing in 8–14 steps (4–7 complete strides) of data per participant-speed.

All spatio-temporal measurements describing walking were extracted from the walkway datausing custom walkway software, and selected variables were identified for subsequent analysis. Fourspecific dependent variables, identified relative to previous literature findings, (velocity, Vel; swing

J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906 901

phase percent of cycle, SW%; double support percent of cycle, DS%; heel to heel stance width, StW)were selected to address Aims 1 and 2, while 14 measures were identified as independent variablesfor the predictive portion, Aim 3, of the study. Bilateral descriptors were retained for the predictiveportion of the analysis owing to reported relationships between gait asymmetry and obesity (Hills& Parker, 1992). Variables used for the respective analyses are specified in Table 1.

2.4. Statistical analysis

An independent t-test was used to test BMI% between groups (NW vs OWO) to determine whethergroup membership was unique. Two-way mixed model Group � Speed analyses of variance (ANO-VAs), with speed as the repeated factor, were used to address study Aim 1. Group � Gender at levelsof Speed ANOVAs were used to address Aim 2. All analyses were conducted for each of the four depen-dent variables (Table 1). Level of significance was set at a = .05.

A stepwise multiple regression procedure was used to predict BMI% from a 14-independent vari-able set (Table 1). Variable entry was set at a = .15 with the overall model significance set ata = .05. This approach of liberal variable entry with stringent overall model significance has beenshown to address the potential confounding multicollinearity issue that can arise from related inde-pendent variables (Dufek & Bates, 1992; Dufek, Mercer, & Griffin, 2009) and eliminates the need ofarbitrary selection between potentially related variables. Three prediction model procedures wereconducted: (1) by speed, (2) by gender, and (3) by gender at levels of speed. All statistical analyseswere conducted using Statistical Analysis Software (Version 9.1; SAS Institute, Inc., Cary, NC).

3. Results

3.1. Participant grouping

Results of the post hoc participant grouping based upon BMI% resulted in 56 members in the NW group(14.7 ± 1.5 yrs; 51.2 ± 9.4 kg; 159.6 ± 10.2 cm) and 55 members in the OWO group (14.9 ± 1.2 yrs;79.8 ± 17.7 kg; 164.8 ± 8.2 cm). There was a significant difference between groups (t(109) =�13.83,p < .001) for BMI% with the NW average (57.9%) less than OWO average (95.5%). There was a near equal

Table 1Variables by study aims.

Dependent/independent variable Text abbreviation

Study Aims 1 and 2Velocity (m/s) VelStance width (cm) StWDouble support (% gait cycle) DS%Swing phase (% gait cycle) SW%

Study Aim 3Left step time (s) LStpRight step time (s) RStpLeft step length (cm) LStLRight step length (cm) RStLStride length (cm) SLStance width (cm) StWLeft swing (% gait cycle) LSW%Right swing (% gait cycle) RSW%Left stance (% gait cycle) LSt%Right stance (% gait cycle) RSt%Left single support (% gait cycle) LSS%Right single support (% gait cycle) RSS%Left double support (% gait cycle) LDS%Right double support (% gait cycle) RDS%

902 J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906

proportion of group membership by gender for the NW (male = 33, female = 23) and OWO (male = 31,female = 24) groups.

3.2. Effects of overweight–obesity and speed (Aim 1)

Descriptive results for both groups at levels of speed are presented in Table 2.The Group � Speed ANOVA results identified a significant, F(1,104) = 4.65, p = .0334, ordinal inter-

action for SW%, with both groups demonstrating increased SW% with increased speed. There were nosignificant Group � Speed interactions observed for Vel (p = .2567), DS% (p = .0565), or StW (p = .2051).As anticipated per study design, a significant speed effect was observed for Vel (F(1,104) = 629.2,p < .0001), DS% (F(1,104) = 273.4, p < .0001), and SW% (F(1,104) = 239.7, p < .0001). There was no sig-nificant speed effect for StW (p = .3467). Velocity was 0.37 m/s greater for fast versus preferred speeds,while DS% was 3.3% longer during preferred speed walking and SW% was 1.7% longer during fast speedwalking. There was a significant group difference for Vel (F(1,104) = 4.55, p = .0352), DS% (F(1,104) =59.13, p < .0001), SW% (F(1,104) = 20.49, p < .0001), and StW (F(1,104) = 20.49, p < .0001). Walking Velwas 0.07 m/s faster and SW% 2.1% longer for NW, while DS% was 4.2% longer and StW 0.025 m greaterfor OWO.

3.3. Effects of overweight–obesity and gender (Aim 2)

Descriptive results for both groups at levels gender are presented in Table 2. There were no significantGroup � Gender interactions observed for any of the dependent variables (p > .05). The only gender dif-ference identified was a significantly greater StW for males, with an average difference = 0.25 cm. Groupdifferences reflected results presented for Aim 1, with all variables significantly different between NWand OWO.

Table 2Mean and standard deviation values for Aims 1 and 2.

NW OWO Total

Female Male Female Male NW OWO(n = 23) (n = 33) (n = 24) (n = 31) (n = 56) (n = 55)

Pref.Vel (m/s) 1.28 1.25 1.18 1.17⁄ 1.25⁄ 1.17#

sd 0.15 0.19 0.16 0.16 0.17 0.16DS% 22.71 23.34 27.37 27.95⁄ 23.80⁄ 8.03#

sd 2.43 3.25 3.50 2.51 3.13 3.19SW% 38.57 38.48 36.37 36.00⁄ 38.16⁄ 35.96�#

sd 1.31 1.75 2.06 1.35 1.71 1.74StW (cm) 7.73 8.40^ 10.14 11.54^⁄ 8.59⁄ 11.09sd 3.40 2.67 2.85 4.10 3.35 2.99

FastVel (m/s) 1.57 1.64 1.53 1.57 1.60⁄ 1.55#

sd 0.17 0.24 0.20 0.22 0.21 0.21DS% 20.28 20.16 23.86 24.02⁄ 20.69⁄ 24.47#

sd 2.43 3.28 3.78 3.07 2.99 3.54SW% 40.03 39.88 38.09 38.11⁄ 39.71⁄ 37.82�#

sd 1.23 1.66 2.03 1.85 1.61 1.91StW (cm) 8.15 8.75^ 10.20 11.21^⁄ 8.80⁄ 11.08sd 2.87 2.59 3.28 3.47 2.94 3.32

Note: sd = standard deviation; a = .05; � = significant Group � Speed interaction; # = significant speed main effect; � = signifi-cant group main effect; ^ = significant gender main effect; Vel = velocity; DS% = double support percent of cycle; SW% = swingphase percent of cycle; StW = heel to heel stance width.

Table 3Summary of prediction of BMI% by speed, gender and gender at levels of speed.

Independent variable Explained variance Graphic representation

By speed: preferredRDS% 41.6RStL 4.8StW 2.3

Total EV 48.7By speed: fastRDS% 26.9LStp 3.7SL 4.5

Total EV 35.1

By gender: femaleRDS% 28.4SL 10.7RStp 4.6

Total EV 43.7By gender: maleLSW% 30.7SL 5.8StW 2.8LDS% 2.6LStp 2.5

Total EV 44.4

By speed and gender: preferred – femaleRDS% 40.8LDS% 11.1LStp 4.4

Total EV 56.3By speed and gender: preferred – maleRDS% 26.3

Total EV 26.3

By speed and gender: fast – femaleLSt% 46.2StW 3.4RSt% 2.2RStL 4.3

Total EV 56.1By speed and gender: fast – maleRDS% 27.3LStp 4.8SL 5.1

Total EV 37.2

Note: Refer to Table 1 for variable abbreviations.

J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906 903

904 J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906

3.4. Predictors of BMI% (Aim 3)

Prediction of the continuous measure of BMI% from spatio-temporal characteristics of walking pro-duced models of weak to moderate strength ranging from 26.3% to 56.3% explained variance (EV)across all models. The prediction results are summarized in Table 3 by variable contribution at eachlevel of contrast. The strongest models (averaging 56.2% EV) were exhibited for females by speed, withDS% being the primary distinguishing characteristic across all models (Table 3).

4. Discussion

The overall purpose of the study, under which the three specific aims fell, was to examine the ef-fects that conditions of overweight and obesity have on gait characteristics of adolescents. This broadpurpose was examined using a variety of approaches in order to more thoroughly expose any relation-ships that may exist. As well, the study was carried out using a large number of participants. A post hocpower analysis indicated statistical power values ranging from 87.7% to 99.9% for the dependent vari-ables assessed in Aims 1 and 2. It is also important to note that while national statistics suggest that20% to 36% of children in the U.S. are OWO (Daniels et al., 2005; Dehghan et al., 2005; Whitaker et al.,2004), the present study sample consisted of nearly 50% membership in the OWO group.

McMillan et al. (2010) performed a kinematic analysis of walking for the same age range popula-tion (12–17 yrs) as in the current study. Due to protocol differences including matching participants inhealthy weight and obese groups based upon preferred walking velocity, direct study comparisonscannot be made. As a result of reported differences in sagittal and frontal plane lower extremity kine-matics and kinetics between groups, the researchers suggested that muscle weakness may be a causeof observed differences. The present study results, which identified longer DS% time for OWO, may bedue to the inability of these individuals to control and accelerate the center of mass over the base ofsupport during the stance phase, perhaps due to lack of strength (Granacher, Gollhofer, & Kriemler,2010). Such conjecture deserves further research attention.

In contrast, Peyrot et al. (2010) found few differences in spatio-temporal characteristics of walkingfor obese adolescents following a weight loss program. Their protocol was unique in that each partic-ipant served as their own baseline/control. It is likely that inherent movement patterns would requiresome time, likely greater than the 12 week period of assessment in this study, to effect change in grosslocomotor function. Despite this limitation, significant differences in stride length were observed. Sin-gle support (%) did not change, and DS% was not reported.

While not specifically focusing on obesity, Granacher et al. (2010) reported that balance trainingsignificantly improved postural control and performance in jump height and leg extensor force devel-opment rate (implied lower extremity strength improvement) in adolescents. The particular focus ofthis work was on potential to reduce lower extremity injury prevalence, which is also an objective inobesity prevention. Whether balance training is appropriate to reduce DS% time during normal walk-ing for OWO adolescents, with the intent of reducing long-term injury potential, is not known. Exam-ination of center of pressure patterns as a measure of dynamic balance and determination ofdifferences between NW and OWO individuals may provide insight into any possible potential bene-fits that balance training might provide for OWO adolescents.

Due to the unbalanced gender demographics in the McMillan et al. (2010) study, the researcherswere unable to assess gender differences; however, it was suggested that future studies explore thisfactor. The present study consisted of 47 females and 64 males. Of these, 24 females and 31 maleswere members of the OWO group. The only gender difference observed between NW and OWO wasStW, with males exhibiting slightly greater (2.5 cm) values. This result may be attributable to the factthat the males were larger physically (height: 1.65 ± 0.09 m, mass: 67.1 ± 19.5 kg) than the females(height: 1.57 ± 0.07 m, mass: 62.6 ± 20.7 kg). Confounding this outcome was the observed gender dif-ference in prediction of BMI% with stronger models generated for females versus males. This outcomemay suggest that males produced a less consistent walking pattern than females, indicating that theremay be a gender difference in walking characteristics of adolescents in general, without regard for adi-posity. Gender differences may also become more pronounced with maturation as this population is

J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906 905

growing to the status of young adults. The 5-year age range among participants (12–17 yrs) resulted ina wide range of physical maturation among participants, allowing only speculation as to thisargument.

The prediction portion of this study was conducted in an attempt to identify any common locomo-tor characteristics that may be related to increased BMI%. Results consistently identified longer DS% asbeing both a salient characteristic of OWO gait as well a strong predictive measure of BMI%. IncreasingDS% may be a compensatory adaptation displayed by OWO adolescents to control inertial character-istics of adipose tissue as it is moved over the base of support during stance. We suggest this may alsobe an adaptive strategy to reduce joint loads as increased DS% is accompanied by a shorter period ofsingle support during walking. In addition, preferred speed walking produced a stronger predictivemodel than fast walking, perhaps because OWO individuals tend to adopt a slower walking velocity.Browning and Kram (2007) found that joint loads became similar between normal weight and obeseadults when participants were allowed to self-select walking speed. Perhaps OWO adolescents seek toaccomplish this same outcome by maximizing DS%. As an extension of this idea, perhaps OWO ado-lescents who seek to lose weight through activity including walking should be coached not to focuson walking faster, but rather to walk longer, thus minimizing acute lower extremity joint loading.

The school setting in which data were obtained for this study provided a familiar environment forthe adolescent participants. At the same time, the non-laboratory setting limited the scope of datawhich could be obtained. Given the differences observed in the present study, it would be helpfulto add to this base of information in the form of joint kinetics and/or three dimensional gait analysisto better understand why OWO adolescents walk differently than their NW peers. Furthermore, suchanalyses may provide insight as to the potential long-term health risks associated with adolescentobesity and locomotion.

5. Conclusion

Child obesity is recognized as an important issue with potential long-term health implications, andtherefore, activity is typically encouraged to reduce BMI in OWO individuals. The biomechanicalimplications of obesity on activities including walking have received some attention in the adult pop-ulation. However, less is known about any relationships between OWO in adolescents and locomotorcharacteristics. Results of this investigation have identified differences in walking Vel, DS%, SW%, andStW for OWO adolescents versus their NW peers, with DS% being most highly predictive of BMI%. Apossibility that such differences in walking patterns at a young age may precipitate to lower extremityjoint dysfunction in adulthood deserves research attention.

Acknowledgments

Funding provided by a University of Nevada, Las Vegas School of Nursing/School of Allied HealthSciences Obesity-Related Award. The study sponsor had no role in the conception or completion of thisresearch. The authors wish to thank Mr. Jeff Kurrus for his editorial comments during manuscriptpreparation.

References

Anderson, J., & Felson, D. (1988). Factors associated with osteoarthritis of the knee in the first national health and nutritionexamination survey (HANES I). American Journal of Epidemiology, 128, 179–189.

Aschemeier, B., Lange, K., Kordonouri, O., & Danne, T. (2008). Pediatric obesity and type 2 diabetes: Strategies for prevention andtreatment. Practical Diabetes International, 25, 368–375.

Barnett, L., Beurden, E. v., Morgan, P., Brooks, L., & Beard, J. (2010). Gender differences in motor skill proficiency from childhoodto adolescence. A longitudinal study. Research Quarterly for Exercise and Sport, 81, 162–170.

Browning, R., & Kram, R. (2007). Effects of obesity on the biomechanics of walking at different speeds. Medicine and Science inSports and Exercise, 39, 1632–1641.

Caprio, S., Daniels, S. R., Drenowski, A., Kaufman, F. R., Palinkas, L. A., Rosenbloom, A. L., et al (2008). Influence of race, ethnicity,and culture on childhood obesity: Implications for prevention and treatment. Diabetes Care, 31, 2211–2221.

Centers for Disease Control & Prevention; CDC. (2009). Defining childhood overweight and obesity. Retrieved March 10, 2009.Available from http://www.cdc.clv/NCCDPHP/DNPA/obesity/childhood/defining.htm.

906 J.S. Dufek et al. / Human Movement Science 31 (2012) 897–906

Daniels, S., Arnett, D., Eckel, R., Gidding, S., Hayman, L., Kumanyika, S., et al (2005). Overweight in children and adolescents:Pathophysiology, consequences, prevention, and treatment. Circulation, 111, 1999–2012.

Dehghan, M., Akhtar-Danesh, N., & Merchant, A. (2005). Childhood obesity, prevalence and prevention. Nutrition Journal, 4.doi:10.1186/1475-2891-4-24.

Devita, P., & Hortobágyi, T. (2003). Obesity is not associated with increased knee joint torque and power during level walking.Journal of Biomechanics, 36, 1355–1362.

Dufek, J. S., & Bates, B. T. (1992). Lower extremity performance models for landing. Human Movement Science, 11, 299–318.Dufek, J. S., Mercer, J. A., & Griffin, J. R. (2009). The effects of speed and surface compliance on shock attenuation characteristics

for male and female runners. Journal of Applied Biomechanics, 25, 219–228.Granacher, U., Gollhofer, A., & Kriemler, S. (2010). Effects of balance training on postural sway, leg extensor strength, and

jumping height in adolescents. Research Quarterly for Exercise and Sport, 81, 245–251.Haubenstricker, J., Wisner, D., Seefeldt, V., & Branta, C. (1997). Gender differences and mixed-longitudinal norms on selected

motor skills for children and youth. Journal of Sport and Exercise Psychology, NASPSPA Abstracts, 19, S63.Hills, A., & Parker, A. (1992). Locomotor characteristics of obese children. Child: Care, Health and Development, 18, 29–34.Lawrence, V., & Kopelman, P. (2004). Medical consequences of obesity. Clinics in Dermatology, 22, 296–302.Malatesta, D., Vismara, L., Menegoni, F., Galli, M., Romei, M., & Capodaglio, P. (2009). Mechanical external work and recovery at

preferred walking speed in obese subjects. Medicine and Science in Sports and Exercise, 41, 426–434.McMillan, A. G., Auman, N. L., Collier, D. N., & Williams, B. (2009). Frontal plane lower extremity biomechanics during walking in

boys who are overweight versus healthy weight. Pediatric Physical Therapy, 21, 187–193.McMillan, A. G., Pulver, A. M. E., Collier, D. N., & Williams, D. S. B. (2010). Sagittal and frontal plane joint mechanics throughout

the stance phase of walking in adolescents who are obese. Gait & Posture, 32, 263–268.Oakley, A., Booth, M., & Chey, T. (2004). Relationship between body composition and fundamental movement skills among

children and adolescents. Research Quarterly for Exercise and Sport, 75, 238–247.Peyrot, N., Jean-Benoit, M., Thivel, D., Isacco, L., Taillardat, M., Belli, A., et al (2010). Mechanical work and metabolic cost of

walking after weight loss in obese adolescents. Medicine and Science in Sports and Exercise, 42, 1914–1922.Shultz, S., Sitler, M., Tierney, R., & Hillstrom, H. (2009). Effects of pediatric obesity on joint kinematics and kinetics during 2

walking cadences. Archives in Physical Medicine and Rehabilitation, 90, 2146–2154.Sorsdahl, A. B., Moe-Nilssen, R., & Strand, L. I. (2008). Test–retest reliability of spatial and temporal gait parameters with cerebral

palsy as measured by and electronic walkway. Gait & Posture, 27, 43–50.Stürmer, T., Günther, K.-P., & Brenner, J. (2000). Obesity, overweight and patterns of osteoarthritis: The Ulm osteoarthritis study.

Journal of Clinical Epidemiology, 53, 307–313.Wearing, S., Hennig, E., Byrne, N., Steele, J., & Hills, A. (2006). The impact of childhood obesity on musculoskeletal form. Obesity

Reviews, 7, 209–218.Whitaker, J. A., Davis, K. L., & Lauer, C. (2004). Is there a need for screening for type 2 diabetes in seventh graders? Journal of the

American Academy of Nurse Practitioners, 16, 496–501.