The effects of small and large varus alignment of the foot-ankle complex on lower limb kinematics and kinetics during walking: A cross-sectional study

Abstract

Background:

The alignment of the foot-ankle complex may influence the kinematics and kinetics of the entire lower limb during walking.

Objectives:

This study investigated the effect of different magnitudes of varus alignment of the foot-ankle complex (small versus large) on the kinematics and kinetics of foot, ankle, knee, and hip in the frontal and transverse planes during walking.

Design:

Cross-sectional study.

Method:

Foot-ankle complex alignment in the frontal plane was measured as the angle between the metatarsal heads and the inferior edge of the examination table, measured with the volunteer in prone maintaining the ankle at 0° in the sagittal plane. The participants (n=28) were divided into two groups according to their alignment angles. The first group had values equal to or inferior to the 45 percentile, and the second group had values equal to or above the 55 percentile. The lower limb kinematics and kinetics were evaluated with the participant walking at self-select speed in an instrumented treadmill.

Results:

The group of large varus alignment showed significantly higher (p<0.03) forefoot inversion angle at initial contact, amplitude of rearfoot-shank eversion, and peak of inversion ankle moment. There were no differences (p>0.05) between the groups for knee and hip amplitudes and moments in the frontal and transverse planes. The durations of rearfoot-shank eversion, knee abduction, knee medial rotation, hip adduction, and hip medial rotation were not different between groups (p>0.05).

Conclusion:

Large varus alignment of the foot-ankle complex may increase the magnitude of foot pronation and ankle inversion moment during walking.

INTRODUCTION

Excessive foot pronation during walking is a commonly observed movement pattern that is characterized by increased magnitude and duration of pronation (Michaud, 1993). One factor that seems to influence foot pronation is a varus alignment of the foot-ankle complex, in which the forefoot, rearfoot, and shank are in an inverted position in non-weight-bearing (Donatelli, 1987; Tiberio, 1988). During weight-bearing activities such as walking, it is suggested that a combination of forefoot, rearfoot, and shank varus alignments may all contribute in varying degrees to forefoot contact with the ground in an excessively inverted position (Monaghan et al., 2014, 2013; Tiberio, 1988). In this paper, we refer to this alignment as the “global” alignment angle of the foot-ankle complex, because it represents all influences acting on the forefoot during ground contact. Thus, this angle determines the torques produced by the ground reaction force (GRF) on the subtalar, midtarsal, and tarsometatarsal joints. With a varus alignment of the foot-ankle complex, the GRF is applied more laterally to the foot (Lugade and Kaufman, 2014), which increases the external eversion moment, and may lead to excessive foot pronation (Tiberio, 1988). Therefore, an excessive varus alignment of forefoot, rearfoot, and shank may influence foot pronation during weight-bearing activities.

Studies have shown an association between varus foot-ankle complex alignment and excessive foot pronation (Buchanan and Davis, 2005; Cruz et al., 2019; Monaghan et al., 2014, 2013; Silva et al., 2014; Souza et al., 2014; Villarroya-Aparicio et al., 2015). Other studies, in contrast, have not found effects of foot varus on foot pronation (Alonso-Vázquez et al., 2009; Cardoso et al., 2020; Cornwall et al., 2004; Donatelli et al., 1999; Hamill et al., 1989; Jarvis et al., 2017; McPoil and Cornwall, 1996). It should be noted that most studies that have shown a significant relationship between foot pronation and varus alignment used the global alignment angle, and not an alignment measure of isolated segments. Moreover, some studies (Cruz et al., 2019; Monaghan et al., 2013) have only found an increase in forefoot eversion and not in rearfoot eversion in individuals with greater global varus alignment, whereas other studies (Souza et al., 2014; Villarroya-Aparicio et al., 2015) observed the effect of global varus alignment on rearfoot eversion. As rearfoot motion in the frontal plane, which is often used to index foot pronation during walking, may be affected by the degree of varus alignment of the shank, rearfoot and forefoot, these inconsistent and contradictory findings need to be clarified.

During closed chain activities, motions of different joints are interdependent. Foot pronation is often coupled with shank, thigh, and hip medial rotation, since the talus transfers the rearfoot motions to the lower limb (Khamis and Yizhar, 2007; Souza et al., 2010). Also, it has been proposed that calcaneal eversion is coupled with knee abduction and hip adduction (Barwick et al., 2012; Resende et al., 2015; Tateuchi et al., 2011). However, there is limited evidence of the effect of varus alignment on the kinematics and kinetics of the entire lower limb during walking, since most studies restricted their analysis to foot kinematics. Studies including knee and hip kinematics were conducted with children (Alonso-Vázquez et al., 2009; Villarroya-Aparicio et al., 2015) or with subjects performing squatting or jumping (Bittencourt et al., 2012; Silva et al., 2015). Knowledge about the impact of excessive global varus alignment on the biomechanics of the entire lower limb during walking would improve understanding of foot mechanics during this activity and the consequences of varus alignment in other joints.

The goal of the present study was to investigate the effect of different magnitudes (small versus large) of global varus alignment of the foot-ankle complex on the kinematics and kinetics of foot, ankle, knee, and hip in the frontal and transverse planes, during the stance phase of walking. The hypothesis was that the group of individuals with large global varus alignment would have greater foot pronation, with its associated proximal lower limb effects, compared to the group with small global varus. Specifically, the effects of larger varus would be evidenced by (i) increased forefoot inversion at initial contact; (ii) increased magnitude and duration of rearfoot eversion relative to shank; (iii) increased excursions magnitude and duration of knee abduction, knee medial rotation, hip adduction and hip medial rotation; and (iv) increased internal moments of ankle inversion, knee abduction, knee lateral rotation, hip abduction, and hip lateral rotation.

METHODS

Participants

This cross-sectional study was approved by University’s Institutional Review Board (n°. 3495E) and included a convenience sample of 36 volunteers. The number of participants was determined based on an estimated large effect size (d=1.2), statistical power of 80% and probability of type I error of 5%. A minimum of 24 subjects (12 participants per group) was needed to detect between-subject differences.

The volunteers selected should have age between 18 and 40 years. The exclusion criteria were: any reported lower extremity or trunk pain or injury in the last seven days; any strikingly visible walking abnormalities (e.g., limping); a history of lower limb or trunk surgery; current use of foot orthoses; and inability to perform any of the procedures required during this study. Given that increased toe-out angle may be related to foot pronation (Rosenbaum, 2013; Wright et al., 1964) and knee moment in the frontal plane (Simic et al., 2013), we excluded participants with a large toe-out angle. As no reference values are available for toe-out classification, we opted to establish the cut-off value as corresponding to the percentile equal or above 80 of our sample (i.e., toe-out angle higher than 17°). Eight participants were excluded due to the presence of toe-out angle higher than the cut-off value. Data from 28 participants (11 men and 17 women) with a mean age of 22.54±4.02 years, mean body mass of 64.38±10.60 kg and mean height of 1.69±0.89 m were analyzed. The diagram representing the flow of participants is shown in Figure 1.

Figure 1 –

Flow diagram displaying the progress of all participants through the study.

Procedures

Each volunteer signed an informed consent form prior to testing.

Foot-ankle complex global alignment

An examiner measured the frontal plane alignment of the foot-ankle complex with the participant in the prone position, the superior aspect of the medial malleolus was aligned with the inferior border of the table, and the hip in neutral abduction/adduction (Figure 2) (Mendonca et al., 2013). A camera stabilized on a tripod was positioned at the lower end of the examination table, an inclinometer was used to level the camera horizontally, and a goniometer was used to align the camera screen parallel to the inferior border of the table. The participant’s lower limb was positioned with the calcaneus facing up. The evaluator attached a metal rod to the metatarsal head using a velcro strap. The examiner positioned the ankle at 0° in the sagittal plane with a goniometer, asked the participant to hold this position actively and took a photo. This last procedure was carried out three times. The global alignment was determined as the angle between the metatarsal head and the inferior border of the examination table (Gross et al., 2007; Monaghan et al., 2013).

Figure 2 –

(A) Measure of global foot-ankle complex alignment of the right lower limb. (B) The angle measured represented with a dashed line. This angle corresponded to the frontal plane angle between the inferior edge of the examination table and the line of the metatarsal heads.

The measurement of global alignment was performed by two experienced examiners who received the same training to perform the measurement. A pilot study with ten participants was conducted to investigate the reliability of this measure. The Intraclass Correlation Coefficient (ICC_3,3) for the intra-examiner reliability was 0.91 and 0.95 for examiners 1 and 2, respectively.

The global alignment of the foot-ankle complex described above was developed to capture the combined effects of bone alignment (i.e., bones from forefoot, rearfoot and shank) and inversion mobility at the midfoot joint complex (Paes et al., 2019; Souza et al., 2014). The result of this measure is influenced by the rearfoot alignment, as the increase in rearfoot varus will also increase metatarsal head inversion angle (Mendonca et al., 2013). The horizontal reference line used, instead of the traditional distal third of the shank bisection, allowed including shank varus/valgus alignment in the measured angle. Finally, by asking the participant to maintain the ankle positioned at 0° in the sagittal plane, the inversion mobility of the midfoot joints influenced the measured angle, as the contraction of the anterior tibialis muscle pulls the midfoot joints into inversion, and the inversion magnitude will depend on the mobility.

Walking biomechanics

The biomechanical evaluation of the lower limbs was performed by capturing the movements of the following segments during walking on the treadmill: foot (analyzed as a single segment as well as segmented into forefoot and rearfoot), shank, thigh, and pelvis. Reflective markers and clusters were placed in the lower limbs (Figure 3).

Figure 3 -.

Set up of reflective markers. Anatomical markers are displayed in black, whereas tracking markers are displayed in white. Markers that are both anatomical and tracking are displayed in black and white. RIC = right iliac crest; LIC = left iliac crest; RASIS = right anterior superior iliac spine; LASIS = left anterior superior iliac spine; SC = sacrum marker placed between the posterior superior iliac spines; RGT = right great trochanter; LGT = left great trochanter; RLK = lateral epicondyle of the right femur; LLK = lateral epicondyle of the left femur; RMK = medial epicondyle of the right femur; LMK = medial epicondyle of the left femur; RLM = right lateral malleolus; LLM = left lateral malleolus; RMM = right medial malleolus; LMM = left medial malleolus; R5MB = right fifth metatarsal base; L5MB = left fifth metatarsal base; R5MH = right fifth metatarsal head; L5MH = left fifth metatarsal head; R2MB = right second metatarsal base; L2MB = left second metatarsal base; R2MH = right second metatarsal head; L2MH = left second metatarsal head; R1MB = right first metatarsal base; L1MB = left first metatarsal base; R1MH = right first metatarsal head; L1MH = left first metatarsal head; RPT = right peroneal tuberosity; LPT = left peroneal tuberosity; RST = right sustentaculum tali; LST = left sustentaculum tali.

First, the walking speed was obtained by asking the participant to walk barefoot, at his/her natural speed, for five trials around a 15m walkway. The walking speed was measured using a mechanized gait carpet (GAITRite Inc., Clifton, USA) positioned on the walkway. The final walking speed was defined as the mean of the five measured repetitions. The speed determined by the walkway was the self-selected speed used during data collection on the treadmill.

Second, a five-second static trial was performed with the participant in standing posture. Third, the participant walked for two minutes on a split-belt instrumented treadmill (Bertec Corporation, Columbus, USA) for familiarization. Fourth, data were collected for 80 seconds with the participant walking on the treadmill at the self-selected speed determined previously. Data were recorded with a Vicon motion analysis system (Oxford Metrics, Oxford, UK). The sample rates of the cameras and force plates were 100 Hz and 1000 Hz, respectively.

Data processing and reduction

Foot-ankle complex global alignment and participant categorization

The right lower limb was chosen for all analyses. The global alignment was determined by analyzing the photos with a routine written in MATLAB® software. The angle formed by the metatarsals heads line and the lower edge of the assessment table was calculated in each photo, and the mean angle of the three photos was used for analysis. The sample was split into two groups. The group with small global varus alignment consisted of participants presenting values less than or equal to the 45th percentile. The group with large varus global alignment was with the percentile greater than or equal to the 55^th percentile.

Walking biomechanics

The kinematic and kinetic data collected during walking was processed with the Visual3D software (C-Motion Inc, Rockville, United States). The biomechanical model of forefoot (Leardini et al., 2007), rearfoot (Leardini et al., 2007), foot (C-Motion Visual3D Documentation, 2015), shank (Bruening et al., 2012; Manal et al., 2000), thigh (Schache et al., 2008; Wu et al., 2002) and pelvis (Wu et al., 2002) were implemented. The marker trajectories and GRF data were filtered using fourth-order Butterworth low-pass filters with 6 Hz and 10 Hz cut-off frequencies, respectively (Robertson and Dowling, 2003). The following angles were obtained: (1) forefoot: movement of the forefoot relative to the laboratory; (2) foot: movement of the foot relative to the laboratory; (3) rearfoot-shank: movement of the rearfoot relative to the shank; (4) knee: movement of the shank relative to the thigh; (5) hip: movement of the thigh relative to the pelvis. The following Cardan sequence was used: medial-lateral, anteroposterior, and superior-inferior axes. The internal moments of the ankle, knee, and hip were calculated using inverse dynamics and were normalized by body mass. Foot initial contact and foot toe-off were determined using the vertical GRF. The strides in which the participant crossed the treadmill midline were excluded.

The angle and moment data were exported from Visual3D and a MATLAB® routine was developed to calculate the following variables: (1) forefoot angle in the frontal plane at initial contact, for which the zero-degree position was considered the position of the forefoot in the static trial; (2) range of rearfoot-shank eversion between initial contact and maximum eversion; (3) duration of rearfoot-shank eversion from initial contact until maximum eversion; (4) range of knee abduction between initial contact and maximum abduction; (5) duration of knee abduction from initial contact until maximum abduction; (6) range of knee medial rotation between initial contact and maximum medial rotation; (7) duration of knee medial rotation from initial contact until maximum medial rotation; (8) range of hip adduction between initial contact and maximum adduction; (9) duration of hip adduction from initial contact until maximum adduction; (10) range of hip medial rotation between initial contact and maximum medial rotation; (11) duration of hip medial rotation from initial contact until maximum medial rotation; (12) maximum internal moment of ankle inversion; (13) maximum knee abduction internal moment during early stance; (14) maximum knee abduction internal moment during late stance; (15) maximum internal moment of knee lateral rotation; (16) maximum internal moment of hip abduction; and (17) maximum internal moment of hip lateral rotation. Furthermore, the toe-out angle was calculated as the angle between the long axis of the foot and laboratory anteroposterior axis. All variables were measured for each stride, and the average across all measured strides in each condition was used for analysis.

Statistical analysis

The normality assumption for data distribution was verified using the Shapiro-Wilke test. Depending on data distribution, independent t-tests or Mann-Whitney tests were used to verify if age, height, body mass, and self-selected walking speed were different between groups. As sex is a categorical variable, the chi-square test was performed to investigate if the frequency distribution of men and women was different between the groups. For the dependent variables, Independent t-tests or Mann-Whitney tests were also used to investigate differences between groups. A probability of type I error (α) of 5% was established.

RESULTS

The small-varus group (n=13) had a mean foot-ankle complex global alignment of 11.54°±4.37° (ranging from 4.64° to 17.32°), whereas the large-varus group (n=13) had a mean global alignment of 22.46°±4.56° (ranging from 19.23° to 26.36°). The statistical analysis demonstrated that these groups were not different (p>0.18) for the sample characteristics (Table 1).

Table 1-.

Descriptive and inferential statistics of the control variables measured for small and large varus groups

Variables	Mean (SD)		Median		Independent t-test or Mann-Whitney or Chi-square test
	Small varus	Large varus	Small varus	Large varus	P	t or U or x2	d or r or rϕ
Age (years)b	22.69 (3.92)	22.46 (4.56)	21.00	21.00	0.69	76.00	0.09
Body mass (kg)a	63.33 (10.39)	66.75 (9.50)	61.92	65.09	0.39	−0.88	0.34
Height (cm)a	1.69 (0.08)	1.69 (0.10)	1.70	1.68	0.97	−0.03	0.01
Sex (men/women)c	6/7	4/9	---	---	0.42	0.65	0.16
Walking speed (m/s)a	1.21 (0.16)	1.13 (0.15)	1.18	1.10	0.18	1.38	0.54

^a= Independent test-t;

^b= Mann-Whitney test;

^c= Chi-square test;

SD = standard deviation; t = statistic for Independent t-test; U = statistic for Mann-Whitney test; x² = statistic for Chi-square test; d = Cohen’s effect size for independent t-test; r = effect size for Mann-Whitney test; rϕ = Phi coefficient (effect size for Chi-square test).

The results showed that the large-varus group had higher values of forefoot inversion angle at initial contact (p=0.02), increased range of rearfoot-shank eversion (p=0.03), and higher peak of ankle inversion moment (p=0.01) compared to the small-varus group (Table 2). There were no other differences between groups. The mean curves of lower limb angles and moments are displayed in Figures 4 and 5.

Table 2-.

Descriptive and inferential statistics of the dependent variables measured for small and large varus groups

Variables	Mean (SD)		Median		Independent t-test or Mann-Whitney
	Small varus	Large varus	Small varus	Large varus	P	t or U	d or r
Forefoot at initial contact in the frontal plane (°)a	2.82 (4.13)	6.72 (3.73)	3.34	6.64	0.02*	−2.52	0.99
Amplitude of rearfoot-shank eversion (°)a	−6.41 (1.90)	−8.26 (2.14)	−6.51	−7.11	0.03*	2.33	0.91
Amplitude of knee abduction (°)a	−6.94 (3.77)	−5.98 (4.32)	−7.33	−6.53	0.55	−0.60	0.23
Amplitude of knee medial rotation (°)a	12.60 (4.68)	9.97 (4.30)	13.09	9.77	0.15	−1.49	0.58
Amplitude of hip adduction (°)a	5.73 (1.80)	6.53 (2.78)	5.81	6.81	0.39	−0.87	0.34
Amplitude of hip medial rotation (°)b	3.07 (3.11)	4.34 (2.29)	2.44	3.68	0.11	65.00	0.32
Duration of rearfoot-shank eversion (%)b	31.61 (5.60)	33.90 (4.61)	33.27	33.94	0.34	156.00	0.20
Duration of knee abduction (%)b	56.19 (16.39	51.55 (19.14)	61.00	61.00	0.96	83.00	0.02
Duration of knee medial rotation (%)b	46.01 (12.62)	49.15 (9.30)	52.62	51.98	0.84	80.50	0.05
Duration of hip adduction (%)a	17.58 (2.57)	16.95 (2.74)	16.87	16.97	0.56	0.60	0.23
Duration of hip medial rotation (%)a	22.13 (14.06)	27.95 (12.60)	21.03	32.35	0.28	−1.11	0.44
Peak of ankle inversion moment (Nm/kg)a	0.21 (0.08)	0.29 (0.08)	0.20	0.32	0.01*	−2.89	1.13
1° peak of knee abduction moment (Nm/kg)a	−0.33 (0.09)	−0.32 (0.08)	−0.31	−0.33	0.74	−0.34	0.13
2° peak of knee abduction moment (Nm/kg)a	−0.33 (0.11)	−0.26(0.09)	−0.31	−0.25	0.08	−1.85	0.73
Peak of knee lateral rotation moment (Nm/kg)a	−0.14 (0.04)	−0.13(0.02)	−0.13	−0.13	0.37	−0.92	0.36
Peak of hip abduction moment (Nm/kg)a	−0.89 (0.12)	−0.88 (0.15)	−0.93	−0.89	0.86	−0.17	0.07
Peak of hip lateral rotation moment (Nm/kg)a	−0.23 (0.12)	−0.27 (0.09)	−0.19	−0.27	0.41	0.84	0.33

^a= Independent test-t;

^b= Mann-Whitney test; SD = standard deviation; t = statistic for Independent t-test; U = statistic for Mann-Whitney; d = Cohen’s effect size; r = effect size for Mann-Whitney* = significant effect. Negative values of angles and moments indicate eversion, abduction or lateral rotation, and positive values indicate inversion, adduct; ion or medial rotation.

Figure 4:

Mean curves of all participants per group for the following angles: (A) rearfoot-shank in the frontal plane; (b) knee in the frontal plane; (c) hip in the frontal plane; (d) knee in the transverse plane and (e) hip in the transverse plane. Negative values of angles indicate eversion, abduction or lateral rotation, and positive values indicate inversion, adduction, or medial rotation.

Figure 5:

Mean curves of all participants per group for the following internal moments: (A) ankle in the frontal plane; (b) knee in the frontal plane; (c) hip in the frontal plane; (d) knee in the transverse plane and (e) hip in the transverse plane. Negative values of moments indicate eversion, abduction or lateral rotation, and positive values indicate inversion, adduction, or medial rotation.

DISCUSSION

As expected, results showed that the large-varus group had higher values of forefoot inversion angle at initial contact, increased range of rearfoot-shank eversion, and higher peak of ankle inversion moment compared to the small-varus group. However, contrary to our hypotheses, no differences between groups were observed for amplitudes and moments of the knee and hip joints. The durations of rearfoot-shank eversion, knee abduction and medial rotation, hip adduction and medial rotation were also not different between groups, which was contrary to our hypotheses. Therefore, the large-varus group had the expected influences at the foot-ankle complex, but no influence on the knee and hip, when compared to the small-varus group.

This study indicated that individuals with higher global varus alignment had greater forefoot inversion angle at initial contact. This result corroborates those of Monaghan et al. (2013) and Monaghan et al. (2014), who reported that global varus alignment predicted forefoot landing angle during walking and running, respectively. The present study also demonstrated that the large-varus group had also increased amplitude of rearfoot-shank eversion. This finding is consistent with those reported by Souza et al. (2014) and Villarroya-Aparicio et al. (2015). However, contrary to our findings, Monaghan et al. (2013) found that individuals with large global varus did not have increased rearfoot eversion. One possible explanation for this result is the difference in values of the global alignment between the studies. Monaghan et al. (2013) established as inclusion criterion an alignment above 15°, and compared groups with moderate (18.5±2.6°) and large varus (28.9±2.9°). The present study compared groups of small (11.54±4.37°) and large varus (22.46°±4 .56°). Thus, perhaps, Monaghan et al. (2013) compared two groups with higher values of global aligment and, as a consequence, both groups may have had increased rearfoot eversion. A recent study (Jarvis et al., 2017) also did not find any association between rearfoot varus alignment and foot kinematics. However, this last study measured only rearfoot alignment according to Root’s method (Root et al., 1971), in which the alignment angle measured is between rearfoot bisection and bisection of the shank distal third with the subtalar joint in the neutral position. Conversely, the angle measured in the present study combines shank, rearfoot, forefoot alignments, and foot mobility.

Contrary to our hypotheses, no difference in duration of rearfoot-shank eversion was found in the present study. Differences in rearfoot-shank eversion duration have been noted during running (Monaghan et al., 2014), but not during walking (Monaghan et al., 2013). Perhaps differences in the duration of rearfoot-shank eversion are more pronounced in activities of higher loads. Consequently, during running, subjects with large varus may be exposed to a greater external ankle eversion moment, leading to a stronger effect on the duration of rearfoot-shank eversion.

The present study showed that the large-varus group presents a higher peak of ankle inversion moment compared with the small-varus group. Again, the large-varus group may be subjected to a more laterally applied GRF, which might increase the external eversion moment. Consequentially, the internal inversion moment may have to increase to compensate for the larger eversion moment arm. The tibialis anterior and tibialis posterior muscles serve to decelerate ankle eversion in early stance (Hunt et al., 2001; Maharaj et al., 2016), and thus we can speculate that these muscles may be some of the structures capable of increasing ankle inversion moment. However, the internal moment calculated through inverse dynamics represents the net moment (Robertson et al., 2014), and it is not possible to deduce which structures are responsible for the increased moment observed. Hsu et al. (2014) found that orthoses with medial forefoot and rearfoot posting increased the internal ankle inversion moment, which may be interpreted as contradictory with the present results. Orthoses, in theory, should simulate a foot with small varus, since the orthosis aims to correct effect of varus. On the other hand, Resende et al. (2015) investigated the ankle moment in individuals wearing pronation inducing sandals and found that increased foot pronation was accompanied by higher internal ankle inversion moment, which is consistent with our findings.

Divergent from our hypothesis, the biomechanics of the knee and hip were not different between the groups. Other studies (Hsu et al., 2014; Resende et al., 2015; Telfer et al., 2013) that induced increases or reductions of foot pronation during walking could identify effects on knee and hip joints. In these studies, the statistical analysis involved paired comparisons, and the same participants were compared before and after the intervention. The paired analysis used in these three studies (Hsu et al., 2014; Resende et al., 2015; Telfer et al., 2013) is statistically more powerful than independent analysis because between-subjects variability is larger than the within-subject variability (Field, 2009). When comparing different subjects, as the present study did, it seems that each subject compensates for excessive foot pronation differently at their proximal joints. This fact increases the sample’s variability and reduces our study power to identify the effects. It is important to point out that both design methods (paired and independent comparisons) are important and correct, but they respond to different questions. So, it is hard to directly compare the results of these studies. Moreover, the lack of differences in the knee and hip may be explained by the values of large varus found in the present study (22.46°±4.56°), which did not include individuals with extremely high values, such as other studies (28.9±2.9°) (Monaghan et al., 2014, 2013). Despite evidence that varus alignment is related to knee and hip injuries (Gross et al., 2007; Lufler et al., 2016; Lun et al., 2004), the present study did not demonstrate effects in the knee and hip joints, which is in accordance with Alonso-Vázquez et al. (2009).

One of the exclusion criteria of the present study was having a large toe-out angle. An additional analysis was done without excluding these participants, and different results were observed. This analysis showed that individuals with greater toe-out had excessive pronation even without the presence of increased global varus alignment, which is consistent with previous research (Rosenbaum, 2013; Wright et al., 1964). Thus, comparisons of the entire sample size (i.e., without excluding participants with large toe-out) showed no differences between the groups with large and small varus. Therefore, this result indicates that the presence of increased toe-out angle is a confounding factor when establishing the relationship between foot pronation and global varus alignment.

Some limitations of the present study must be recognized. The small number of subjects per group was enough to detect only differences of large effect size. Some variables (i.e. amplitude of knee medial rotation and second peak of knee abduction moment) showed moderate effect size and statistical power. Statistical differences between the groups could have been found in these variables if the sample size increases. No measures of muscles electromyography activity were performed. These measures might have been helpful to reveal the source of the observed increase in ankle inversion moment. In addition, we evaluated walking biomechanics on a treadmill, which may be different from overground walking.

The results of this study indicated that individuals with large global varus alignment show increased forefoot inversion at initial contact, increased rearfoot-shank eversion, and greater ankle inversion moment during walking. Therefore, the contribution of global varus alignment to foot pronation was established. However, the effects observed in this study were restricted to the foot mechanics, as no difference between groups was observed in knee and hip joints. When evaluating and treating excessive foot pronation, foot-ankle complex global alignment should be considered as a potential cause of this movement dysfunction. Considering that foot pronation has multiple causes, discovering its source is relevant for clinical practice. If a large varus is found in the assessment, it may be one of the causes for the presence of excessive foot pronation, and may be a target of a clinical intervention.

Highlights.

• Large varus of the foot-ankle complex increased foot pronation during walking

• Large varus alignment increased forefoot inversion at initial contact

• Large-varus group had higher ankle inversion moment compared to small-varus group

• No differences between the groups were observed for the knee and hip joints