Skip to main content

Table 4 Summary of XcoM and MoS definitions and calculations. For case-control studies that showed a statistically significant difference and where data was available, Glass’s Δ is reported as a measure of effect size

From: Use of the margin of stability to quantify stability in pathologic gait – a qualitative systematic review

Paper

CoM definition

Pendulum length

BoS definition

MoS calculation

MoS reference edge

Point of gait

Results as reported in original paper

Standardised results interpretation

Stroke (n = 8)

 Hak, et al. (2013) [14]

Centre of the polygon described by 4 pelvic markers

Maximal height of the origin of the pelvis

AP: Lateral malleolar marker of the leading foot

BoS – XcoM

AP: Posterior

AP: Heel strike

No significant group effects.

No significant difference for MoSML or MoSAP.

ML: Lateral malleolar marker of the leading foot

ML: Lateral

ML: Minimum value per step

 Kao, et al. (2014) [27]

Not specified.

Height of the COM during quiet standing

AP: Front toe marker of the leading foot

BoS - XcoM

AP: Anterior

AP: Heel strike

Post-stroke individuals had smaller average MOSAP (p = 0.042) but no difference in MOSML, compared to controls.

Post-stroke individuals had greater variability of MOSAP and MOSML compared to controls (p < 0.001).

Post-stroke individuals had significantly less stable MoSAP and no difference for MoSML.

Post-stroke individuals had greater MoSAP and MoSML variability, compared to controls. MoSML variability was significantly greater for the affected leg in post-stroke individuals.

ML: Lateral toe marker of the leading foot

ML: Lateral

ML: Heel strike

 Hak, et al. (2015) [23]

Centre of the polygon described by 4 pelvic markers

Maximal height of the estimated CoM

AP: Heel marker of the leading foot

XcoM - BoS

AP: Posterior

AP: Minimum value per step

MoSAP increased when stride length (p < 0.001) and stride frequency (p < 0.001) were increased. MoSML increased when stride frequency was increased (p < 0.001).

MoSAP increased with increased stride length and stride frequency and MoSML increased with stride frequency. Increased MoSAP and MoSML was limited during faster than comfortable stride frequency suggesting inability of post-stroke individuals to regulate MoS using stride frequency.

ML: Lateral malleolar marker of the leading foot

ML: Lateral

ML: Minimum value per step

 van Meulen, et al. (2016) [20]

Fusion of low-pass filtered CoP data with high-pass filtered double-integrated CoM acceleration data.

Vertical CoM position

AP: Midpoint between the front of each foot

MoSAP = XcoM – BoS

MoSML = |BoS – XCoM|

AP: Anterior

AP: Continuous during double-limb support

A positive, significant correlation was found between fall risk and percentage of time spent with a positive MoSAP (r = 0.75, p = 0.014). MoSML asymmetry was not correlated with participant’s fall risk.

MoSAP was more often more stable for more affected post-stroke participants. MoSAP and MoSML were less stable on participants’ affected side.

ML: Lateral shoe position

ML: Lateral

ML: Continuous during double-limb support

 van Meulen, et al. (2016) [19]

Fusion of low-pass filtered CoP data with high-pass filtered double-integrated CoM acceleration data.

Greater trochanter height estimated from total body height

AP: Line connecting the front of each foot.

XcoM-BoS

AP: Anterior

AP: Continuous during double-limb support

Participants with lower BBS scores tend to have a slower walking speed and small positive average MoSAP. There is no significant correlation between BBS and MoSAP (r = 0.41, p = 0.167).

MoSAP was not significantly correlated with a standard clinical parameter, but MoSAP was more often stable for more affected post-stroke participants.

ML: n/a

ML: n/a

ML: n/a

 Vistamehr, et al. (2016) [26]

Cumulative anthropometric segmental mass properties (13 segment)

1.34 x leg length (m)

AP: n/a

BoS - XcoM

AP: n/a

AP: n/a

MoSML was inversely correlated with the clinical scores (BBS and DGI).

MoSML was significantly moderately negatively correlated with other balance measures (more stable for lower Berg Balance Score). When feet were separated, only the affected side correlated with other balance measures.

ML: CoP

ML: Lateral

ML: Heel strike

 Punt, et al. (2017) [24]

Cumulative anthropometric segmental mass properties (14 segment)

Not specified

AP: Not specified

BoS - XcoM

AP: Anterior

AP: Heel strike

MoSAP and MoSML were similar during steady-state gait at a fixed speed for faller and non-faller groups.

MoSAP and MoSML was not significantly different between faller and non-faller groups for the paretic and non-paretic legs.

MoSAP variability was significantly different between faller and non-faller groups for the paretic leg, and for MoSML variability for the paretic and non-paretic leg.

ML: Not specified

ML: Lateral

ML: Heel strike

 Tisserand, et al. (2018) [25]

Cumulative anthropometric segmental mass properties (number of segments not specified)

1.34 x leg length (m)

AP: n/a

BoS - XcoM

AP: n/a

AP: n/a

Post-stroke participants had a larger MoSML than controls during baseline treadmill walking (p < 0.01), with a larger MoSML on the non-paretic side than on the paretic side at ipsilateral foot-strike (p < 0.05).

MoSML was significantly more stable for non-paretic and paretic limbs at heel strike compared to controls. MoSML was significantly more stable for the non-paretic limb compared to the paretic limb at heel strike.

ML: Midpoint between the heel marker and 2nd metatarsal marker

ML: Lateral

ML: Heel strike & toe off

Unilateral transtibial amputees (n = 5)

 Curtze, et al. (2011) [33]

Not specified

1.34 x leg length (m)

AP: n/a

BoS - XcoM

AP: n/a

AP: n/a

There was no significant difference between amputee and control groups for MoSML (p = .763).

MoSML was not statistically different between amputees and controls, or between prosthetic and sound limbs for the amputee group.

ML: AP axis defined by the 2nd metatarsal and calcaneal markers

ML: Lateral

ML: Minimum value during stance phase

 Gates, et al. (2013) [32]

Cumulative anthropometric segmental mass properties (number of segments not specified)

1.34 x leg length (m)

AP: n/a

BoS – XcoM

AP: n/a

AP: n/a

Amputees had a greater mean MoSML than controls (p = 0.018). Amputees had a smaller MoSML on their prosthetic limb compared to intact limb (p = 0.036), while controls had no significant between-limb differences.

MoSML was significantly more stable for amputees than controls.

Amputees had a significantly less stable MoSML on their prosthetic limb compared to their sound limb.

ML: 5th metatarsal marker

ML: Lateral

ML: Minimum value during stance phase

 Hak, et al. (2013) [31]

Centre of the polygon described by 4 pelvic markers

Maximal height of the origin of the pelvis

AP: Lateral malleolar marker of the leading foot

BoS - XcoM

AP: Posterior

AP: Continuous

MoSAP was smaller for amputees than for controls (p = 0.02). In Amputees had a larger MoSML than controls (p = .013).

MoSAP was significantly less stable for amputees than controls.

MoSML was significantly more stable for amputees than controls, possibly due to a compensatory wider step width.

ML: Lateral malleolar marker of the leading foot

ML: Lateral

ML: Continuous

 Beltran, et al. (2014) [30]

Cumulative anthropometric segmental mass properties (13 segment model)

1.34 x leg length (m)

AP: n/a

XcoM – BoS

AP: n/a

AP: n/a

There was no significant difference between mean MoSML and MoSML variability between amputees and controls or between intact and prosthetic limbs for the amputee group.

MoSML was not significantly different between amputees and controls. MoSML variability was not significantly different between amputees and controls.

ML: 5th metatarsal marker

ML: Lateral

ML: Minimum value during stance phase

 Hak, et al. (2014) [28]

Centre of the polygon described by 4 pelvic markers

Maximal height of the estimated CoM

AP: Lateral malleolar marker of the leading foot

XcoM - BoS

AP: Posterior

AP: Heel strike & toe off

The MoSAP was significantly larger (p = 0.018) for the sound limb compared to the prosthetic limb. There was a significant (p = 0.001) increase of MoSAP between initial contact and contralateral toe-off.

MoSAP was significantly more stable at heel strike for the prosthetic limb, compared to the sound limb of amputees, but not significantly different at toe off.

ML: n/a

ML: n/a

ML: n/a

Other amputees (n = 4)

 Hof, et al. (2007) [34]

Low-pass filter of CoP data

1.34 x trochanter height (m)

AP: n/a

BoS - XCoM

AP: n/a

AP: n/a

In amputees MoSML for the prosthetic leg was always larger than for the non-prosthetic leg and larger than the values for the control subjects.

MoSML was significantly more stable for amputees compared to controls at all speeds (Glass’s Δ: control vs. prosthetic limb = 1.6; control vs. non-prosthetic limb = 0.3). MoSML was significantly more stable for amputee’s prosthetic limb compared to their sound limb at all speeds.

ML: CoP

ML: Lateral

ML: Heel strike

 Major, et al. (2013) [3]

Cumulative anthropometric segmental mass properties (number of segments not specified)

Not specified

AP: n/a

BoS - XcoM

AP: n/a

AP: n/a

Amputee step widths were greater than controls at all speeds and prosthetic type (p = 0.002). The XcoM exceeded the lateral borders of the BoS in all amputees at fast walk and when using the prosthetic with greater ankle joint motion, but this never happened in controls.

XcoM frequently exceeded the BoS (became unstable) in the ML direction for the prosthetic group wearing a prosthetic limb with additional ankle motion compared to controls and the same participants wearing a prosthetic limb with more limited ankle motion where the XcoM was always maintained within the BoS (remained stable).

ML: CoP of the stance limb

ML: Lateral

ML: Peak XcoM

 Brandt, et al. (2019) [29]

Cumulative anthropometric segmental mass properties (number of segments not specified)

1.34 x leg length (m) which was the average of the 2 trochanters

AP: n/a

BoS – XcoM

AP: n/a

AP: n/a

Mean MoSML was 5.71 cm (1.18 cm) for the prosthetic limb and 4.92 cm (1.18 cm) for the sound limb during baseline treadmill walking.

MoSML stability was more stable for the prosthetic side compared to the intact side, but this was not compared statistically.

ML: CoP

ML: Lateral

ML: Minimum value per step

Major, et al. (2019) [35]

Cumulative anthropometric segmental mass properties (12 segment)

1.34 x trochanter height (m)

AP: n/a

BoS – XcoM

AP: n/a

AP: n/a

MOSML was significantly greater on the sound limb side compared to the prosthetic limb side (p = 0.005).

MoSML was significantly less stable for the prosthetic limb compared to the sound limb in all conditions.

ML: Fifth metatarsal of the stance limb

ML: Lateral

ML: Minimum value per step

Spinal cord injury (SCI) (n = 2)

 Day, et al. (2012) [38]

Cumulative anthropometric segmental mass properties (13 segment model)

Not specified

AP: n/a

BoS - XcoM

AP: n/a

AP: n/a

Participants with SCI had significantly greater MoSML variability compared to controls (p < 0.007).

MoSML had significantly greater variability in post-SCI participants compared to controls suggesting compensatory control mechanisms to avoid falls.

ML: CoP

ML: Lateral

ML: Minimum value during double-limb support

 Arora, et al. (2019) [13]

Cumulative anthropometric segmental mass properties (12 segment)

Not specified

AP: Anterior foot boundary

BoS - XcoM

AP: Anterior

AP: Heel strike

MoSAP for participants with spinal cord injury was significantly smaller than controls walking at matched speeds (p < 0.01).

MoSAP was not significantly different between SCI participants and controls walking at their self-selected speed. MoSAP was significantly less stable for SCI participants compared to controls walking slower than their self-selected pace to more closely match walking speed on the SCI individuals (Glass’s Δ = 2.9).

ML: n/a

ML: n/a

ML: n/a

Multiple Sclerosis (MS) (n = 2)

 Peebles, et al. (2016) [12]

Geometric centre of the triangle formed by 2 anterior superior iliac spine markers and the midpoint between the 2 posterior superior iliac spine markers

Distance between the estimated CoM and the ankle marker

AP: Toe marker

BoS – XcoM

AP: Anterior

AP: Heel strike & mid-stance

MS participants with gait impairments had a higher MoSAP than controls (p < 0.001) and MS participants without gait impairments (p < 0.001) at heel strike and mid-stance. At heel strike, MS participants with gait impairments had a higher MoSML than controls (p = 0.010).

MoSAP was significantly more stable for MS participants with a gait impairment, compared to those without and compared to controls at heel strike and mid-stance (Glass’s Δ: Heel strike = 1.3; Mid-stance = 1.2). MoSML was significantly more stable for the MS participants with a gait impairment compared to controls at heel strike (Glass’s Δ = 1).

ML: Lateral metatarsophalangeal joint

ML: Lateral

ML: Heel strike & mid-stance

 Peebles, et al. (2017) [39]

Centre of the polygon described by 4 pelvic markers.

Distance between the estimated CoM and the ankle marker

AP: Toe marker

BoS - XcoM

AP: Anterior

AP: Heel strike

MS fallers had a decreased mean MoSAP (p < 0.001) and an increased MoSAP variability (p < 0.001) compared to both MS non-fallers and controls. MS non-fallers had an increased mean MoSML (p = 0.011) compared to controls only, and MS fallers had an increased MoSML variability (p < 0.001) compared to both MS non-fallers and controls.

MS fallers had less stable MoSAP (Glass’s Δ = 1.5) and increased MoSAP variability compared to MS non-fallers and controls. MS non-fallers were more stable for MoSML (Glass’s Δ = 0.6) and had increased MoSML variability compared to controls. MS fallers had increased MoSML variability compared to MS non-fallers and controls.

ML: Lateral metatarsophalangeal joint

ML: Lateral

ML: Heel strike

Parkinson’s Disease (PD) (n = 3)

 Stegemöller, et al. (2012) [37]

Cumulative anthropometric segmental mass properties (15 segment)

Distance between the COM and the centre of the ankle joint in the sagittal plane

AP: Marker positions on the foot

BoS - XcoM

AP: Anterior

AP: Heel strike & toe off

At heel strike and toe from the leading and trailing foot the PD group had a significantly larger MoSAP than controls.

PD participants were more stable than controls during baseline walking at heel strike and toe off for the leading (Glass’s Δ: Heel strike = 6.9; Toe off = 2.6) and trailing (Glass’s Δ: Heel strike = 8.5; Toe off = 5.3) foot in the AP direction.

ML: n/a

ML: n/a

ML: n/a

 Catalá, et al. (2016) [36]

Cumulative anthropometric segmental mass properties (number of segments not specified)

Distance between the estimated CoM and the ankle marker

AP: AP distance between the toes of the anterior foot and heel of the posterior foot

BoS - XcoM

AP: Anterior

AP: Heel strike

MoSAP was significantly lower (p < 0.05) in the PD group compared to controls, reflecting more unstable gait patterns in unperturbed walking at the same walking velocity.

MoSAP was significantly less stable for PD participants compared to controls.

ML n/a

ML: n/a

ML: n/a

 Martelli, et al. (2017) [11]

Cumulative anthropometric segmental mass properties (13 segment)

Instantaneous distance between the body CoM and the ankle joint of the leading leg

AP: Tip of the toe

BoS – XcoM

AP: Anterior

AP: Heel strike

PD participants always walked with a significantly lower MoSAP than controls (p = 0.044). No significant difference for MoSML.

MoSAP was significantly less stable for PD participants than controls. No significant difference for MoSML.

ML: Fifth metatarsal marker

ML: Lateral

ML: Heel strike

Miscellaneous (n = 7)

 McCrum, et al. (2014) [18]

Cumulative anthropometric segmental mass properties (12 segment model)

Instantaneous distance between the body CoM and the ankle joint of the leading leg

AP: Toe marker of the leading foot

BoS - XcoM

AP: Anterior

AP: Heel strike

No significant differences in MoSAP between UPVD patients and controls. Both groups had a positive MoSAP, which indicates a stable body position.

No significant difference for MoSAP between UPVD participants and controls.

ML: n/a

ML: n/a

ML: n/a

 Hoogkamer, et al. (2015) [44]

Cluster of markers at pelvis

Not specified

AP: n/a

BoS - XcoM

AP: n/a

AP: n/a

No significant different between cerebellar lesion participants and controls for MoSML during treadmill walking.

MoSML was not significantly different between cerebellar lesion participants and controls during treadmill walking.

ML: Lateral boundary of the feet

ML: Lateral

ML: Contralateral toe off

 Rijken, et al. (2015) [40]

Cumulative anthropometric segmental mass properties (12 segment model)

0.55 x body height (m)

AP: Heel marker

BoS – XcoM

AP: Anterior

AP: Heel strike

No significant differences between groups were found in MoSAP or MoSML values.

No difference in MoSAP or MoSML for affected participants between mild and moderate severity groups or compared to controls.

ML: Ankle marker

ML: Lateral

ML: Minimum value during stance phase

 Ghomian, et al. (2017) [43]

Cumulative anthropometric segmental mass properties (15 segment)

Distance between the COM and the lateral heel marker at heel strike

AP: Toe marker

BoS – XcoM

AP: Anterior

AP: Heel strike

MoSAP was significantly different for barefoot condition compared to all three shoe conditions. The barefoot condition had a positive MoSAP while all shoe conditions were negative. No significant differences for mean MoSML across all conditions.

MoSAP was significantly different for barefoot compared to all shoe trials. MoSAP was more stable for barefoot than all rocker shoes.

No significant differences were found between any condition for MoSML.

ML: Lateral rocker or 5th metatarsal marker

ML: Lateral

ML: Heel strike

 Simon, et al. (2017) [42]

Cumulative anthropometric segmental mass properties (13 segment)

Trochanteric height (calculation not specified)

AP: n/a

BoS - XcoM

AP: n/a

AP: n/a

MoSML was smaller in the spinal deformity group compared to the control group. 14 spinal deformity participants were unstable and the remaining 3 patients were stable.

MoSML was less stable for spinal deformity participants than controls.

ML: Lateral aspect of the foot created by the line between the lateral toe and lateral malleolar marker

ML: Lateral

ML: Heel strike

 van Vugt, et al. (2019) [41]

Cumulative anthropometric segmental mass properties (number of segments not specified)

Vertical distance between the CoP and the CoM during static trial

AP: Metatarsal marker of the stance foot

BoS - XCoM

AP: Anterior

AP: Heel strike & mid-stance

HSP participants had a significantly lower MoSML at heel strike and mid-stance. HSP participants had a significantly less negative MoSAP at mid-stance, but there was no difference for MoSAP at heel strike.

HSP participants were significantly more stable than controls for MoSAP at mid-stance (Glass’s Δ = 2.1). HSP participants were significantly less stable than controls for MoSML at heel strike and mid-stance (Glass’s Δ: Heel strike = 1.7; Mid-stance = 1.8) .

ML: 2 cm lateral to the 2nd metatarsal marker

ML: Lateral

ML: Heel strike & mid-stance

 de Jong, et al. (2020) [45]

Centre of the polygon described by 4 pelvic markers.

Maximum height of the CoM

AP: n/a

BoS – XcoM

AP: n/a

AP: n/a

No significant difference for MoSML between spinal deformity and control groups.

No significant difference between groups for MoSML.

ML: CoP

ML: Lateral

ML: Minimum value at the start of single-support phase