Effects of deep knee flexion on skin pressure profile with lower limb device: A computational study

Abstract

Knee flexion behavior alters the contact pressure distribution exerted by compression devices during exercise. This study aimed to develop a three-dimensional dynamic finite element model of the lower limb with detailed bony structures, wearing a compression device with higher pressure over the calf, and then to quantify and compare the garment–body interface contact pressure and the cross-section pressure gradient deviation in standing and deep knee flexion postures (30°, 60°, 90°, and 120° of knee flexion). Contact pressure experiment on seven muscle points was applied to validate the model. The cross-section pressure gradient deviation was calculated on landmarks based on deviation along the four axial pathways from the average cross-section pressure gradients. In general, the results demonstrated that the whole pressure profile gradually decreased from the ankle to the thigh with higher compression on the calf in a standing position. Cross-section pressure gradient deviation resulted in a dramatic increase of ∼100% and ∼110% on positions B1 and D on the anterior of calf at 60° flexion, respectively, which resembled an M shape. This phenomenon was caused by the combination of the stretch of clothing during knee flexion, high compression over the calf, and the shape of the lower limb. This finite element model and its findings together could help us to understand the compression effects of sports lower limb devices and garments to enhance walking and running performance, and help to improve the design concepts.

Keywords

compression sports devices deep knee flexion contact pressure pressure deviation textile biomechanics

Introduction

Compression sports lower limb devices and garments are commonly thought to have the potential to enhance performance, decrease discomfort, and reduce the risk of injury in individuals during exercise or recovery. The extrinsic mechanical pressure exerted by compression garments plays a crucial role in increasing cutaneous and subcutaneous interstitial pressure in the lower limb, thereby having positive effects on supporting calf muscle pump, reducing venous hypertension and reflux, improving tissue microcirculation, and minimizing muscle oscillation and muscular soreness.^1–4 In clinical practice, graduated compression therapy, which provides a decreasing pressure from distal to proximal on the lower limb using graduated compression hosiery or bandages, is generally considered as the “gold standard” in the conservative treatment of venous disease. Recent studies indicate that “negative gradient” compression, which provides high compression pressure over the calf, is more effective than graduated compression during walking.⁵ In sports, the supporting effects of devices with varying pressure profiles has been assumed to benefit the lower limb, even with light compression. In endurance exercise, compression tights that applied 6 mmHg and 14.7 mmHg at ankle and calf point in vivo, respectively, improved recovery for performance.⁶ However, there is no general consensus on the optimum pressure profile of compression sports devices or garments to date. In this study, the original design of lower limb devices, which may include but are not limited to compression sports tights, was evaluated using finite element modeling (FEM) to provide graduated compression on the lower limb with calf muscle pumping to enhance walking or running performance. The compression profile of the lower limb device is the combination of the graduated compression distribution and high compression over the calf.

The development of compression devices or garments that can provide benefits for exercise is challenging because of the fundamentally different biomechanics of different body postures in sports and daily life. Keeping knee flexion in good condition at critical phases of the running or walking cycle is one of the key aspects of running or walking well. The use of FEM to calculate the dynamic interface pressure on the human body exerted by compression garments could facilitate the design process by improving our understanding of the dynamic effects of knee flexion on pressure profiles. Recent studies show that the none of the average in-vivo resting pressure at the ankle point of sport compression stockings or sports tights achieved the pressure level of medical compression class I,^7,8 while the tested sport compression stockings, sports tights, and sports leggings exhibited a pressure increase from ankle to calf in vivo.^7–10 In these studies, the pressure sensors were positioned at locations that were not described in the graduated compression stockings normalization.^7–9 Investigating the pressure distribution underneath compression sports devices and garments on the areas of interest during exercise is vitally important to understand its effects on performance. In the current study, seven important muscles were selected in the validation experiment, because they are flexed or extended with knee flexion. For the landmarks along four sides on seven cross sections, the locations were those commonly chosen to analyze the pressure under compression garments.^12,14,18

Several studies have indicated that the pressure profile exerted by compression garments can be influenced by posture or movement. It has been reported that stockings that showed a good graduated compression profile in the standing position may fall outside the ideal band to some degree at the knee-flexed sitting position.¹¹ It has also been observed that the pressure around the ankle and knee was often drastically raised as the joint flexed.¹¹ Recent studies show that changes in joint movement have more significant influences on skin pressure than muscle contraction.¹² The dynamic pressure traces on different locations may not change coincidentally during movement.¹³ Note that only a few studies have tried to investigate the pressure with compression stockings or garments during walking, knee bending, or activities.^{11,12,14–16} When axial gradient, as well as positive von Mises stress gradient, was introduced, the negative axial gradient of interface pressure consistently led to positive axial gradients in adipose tissues on the arm.¹⁷ However, the change of pressure gradient exerted by compression garments or devices is seldom investigated. Information on dynamic compression effects during deep knee flexion is useful in enhancing biomechanical knowledge of the garment–body interaction for researchers and manufacturers involved in the design of compression garments. In this study, the cross-section pressure gradient deviation was introduced to calculate the deviation of pressure on the selected cross section along the axial pathways of the lower limb during deep knee flexion.

Finite element (FE) analysis is widely used in bioengineering in these days, as it allows complicated modeling with nonlinear materials and geometric shapes, irregular boundary conditions, and multi-component assembly.^18–23 It is particularly suitable for modeling the complicated behavior of a human lower limb wearing compression sports devices and garments and then flexing the knee joint. Some FE studies focus on the assessment of contact pressure distribution of the whole leg exerted by compression sportswear or stockings in a resting position.^18,19 The local curvature of the human body surface, and the basic anatomic bony and muscular structures, may be the reason for the large variations of local pressures along the cross section of the lower limb in a standing position.¹⁸ However, only the standing position model without knee flexion behavior has been investigated in previous studies. As far as human activities are concerned, the contact pressure distributions induced by compression garments during knee flexion behavior should be evaluated.

The goal of this study was threefold in that we aimed at developing: (i) a three-dimensional (3D) dynamic FEM for deep knee flexion behavior of the lower limb with details of a human knee joint wearing a compression sports device; (ii) experimental validation on seven selected muscle points on male athletes: tibialis anterior, gastrocnemius medialis, gastrocnemius lateralis, vastus medialis, vastus lateralis, semimembranosus, and rectus femoris; as well as (iii) quantify and compare the contact pressure, and cross-section pressure gradient deviation of the landmarks along four sides (front, medial, back, and lateral sides) on seven cross sections (B: ankle; B1: area at which the Achilles tendon changes into the calf muscles; C: calf at its maximum girth; D: just below the tibial tuberosity; LT: lower thigh; MT: mid thigh; UT: upper thigh) of the lower limb in standing and knee flexion posture.

Methods

Geometry, materials, and mesh of lower limb for FEM

Since the present study focused on contact pressure distribution on the human body surface, the geometry of the detailed anatomy of the knee joint as well as the soft tissue of the lower limb was taken from the reconstruction of geometrical shapes of a commercial 3D anatomic male skin and skeleton model (Zygote Media Group, Inc.). To make a solid 3D representation, the whole male model, including skin, skeleton, and connective tissue, was scaled to a height of 174 cm, which is within the height range of the experimental subjects. The process of reconstruction of the human lower limb consisted of the reconstruction of the hollow leg, femur, tibia, patella, and fibula using the ScanTo3D package in SolidWorks 2008 software (Dassault Systèmes SolidWorks Corp.), as shown in Figure 1. The reconstruction process of the hollow leg was reported in our previous study,^18,19 while the bony structure of the knee joint was presented in the 2011 study by Lin.²⁴ Given the fact that the interface pressure is the main concern during knee flexion in this study, the ligaments and tendon of the knee joint were not involved in the model for simplification.

Figure 1.

Geometric model of compression sports device and lower limb with bony structure.

The parameters of the hyperelastic model of the soft tissue were selected according to the reported data on intact living muscle.²⁵ In that study, the nonlinear elastic behavior of the passive muscle of rat in the transverse direction was described by means of the first-order Ogden model, and the strain energy function of this is shown as follows,^25,26

W = \frac{μ}{α} (λ_{1}^{α} + λ_{2}^{α} + λ_{3}^{α} - 3)

(1)

where λ_i are the principal stretch ratios, and µ and α are the material parameters. The input parameter of the elastic parameters, μ and α, are 15.6 kPa and 21.4, and the material is incompressible.

The parameters of the hyperelastic model of the skin were selected according to those reported by Barbarino et al.²⁷ In that study, the nonlinear elastic behavior of the superficial skin was described by means of the second-order reduced polynomial with the following strain energy function:

W = C_{10} ({\bar{I}}_{1} - 3) + C_{20} ({\bar{I}}_{2} - 3)^{2}

(2)

where C₁₀ and C₂₀ are material parameters. The input parameter of the elastic parameters C₁₀ and C₂₀ were 3.18 kPa and 14.5 kPa, respectively.

{\bar{I}}_{1}

and

{\bar{I}}_{2}

are measures of the distortion in the material.

FE mesh models with 3D tetrahedral elements were built using auto-meshing technique. The size of elements of general instance assigned to the femur, tibia, fibula, and patella were 18, 12, 18, and 2.5 mm, respectively.

Geometry, materials and mesh of compression sports device

The geometry (Figure 1) and material of the compression sports device were developed according to a real compression sports device with polypropylene material and plain knitted structure. The device provides decreasing pressure from distal to proximal (B, C, LT, MT, UT) on the lower limb with negative graduated compression from B1 to C on the calf part. To ensure the flexibility of the compression device, relatively low pressure was exerted on the D part. A third-order Ogden strain model was used to represent the nonlinear elastic incompressive fabric, by fitting the mechanical experimental data with the strain potential energy functions that are built in to Abaqus, which was verified and described in detail in our previous paper¹⁸:

W = \sum_{i = 1}^{3} \frac{2 μ_{i}}{α_{i}^{2}} ({\bar{λ}}_{1}^{α_{i}} + {\bar{λ}}_{2}^{α_{i}} + {\bar{λ}}_{3}^{α_{i}} - 3)

(3)

where µ_i and α_i are material parameters. The input parameters µ₁, µ₂, and µ₃, are −5.931 MPa, 3.107 Mpa, and 2.831 MPa, parameters α₁, α₂, and α₃, are 1.960, 2.212, and 1.708, respectively. The thickness of the fabric is 0.793 mm. The measurement of thickness followed the ASTM D1777 standard. The contact pressure was 4.14 kPa. The temperature was 21℃ and relative humidity 65%. The friction coefficient was 0.4638. The measurement of friction coefficient was performed using the Kawabata Evaluation System for fabric. The eight-node linear solid element was used for the compression sports device structure.

Load and boundary conditions

The load definition during the simulation is divided into two steps: the wearing process and the knee flexion process. During the wearing process, the hollow shell of the soft tissue and bony structure were fixed, while the upper thigh and ankle of the sports device were moving along the leg, which was similar to our previous study.¹⁸ An elastic band which is stiffer than the sports device was designed at the upper thigh of the sports device; it would be tightened up after the wearing process to maintain the normal wearing state of the sports device at the next step.

The second step was to implement the motion of knee flexion. This presented a great challenge since the mechanism of knee flexion remains unclear. Several simplifying assumptions were made with regard to the load and boundary conditions. The inertia load effect was ignored in the simulation. The bony structures are assumed to be rigid bodies. The movement of the knee is decomposed into three major components: rotation about the flexion-extension axis in the posterior femoral condyles, rotation about the longitudinal axis in the tibia, and translation of the femoral condyle.²⁸ The four constraints of passive patellofemoral joint (PFJ) kinematics were depicted by PFJ flexion, PFJ tilt, PFJ translation along femoral axis, and proximodistal axis, according to the literature.²⁹

Experimental validation

Five male distance runners in regular training were recruited for this study (mean age 20.9 ± 1.1 years, height 174.6 ± 6.3 cm, body mass 62.7 ± 8.1 kg, and body mass index (BMI) 20.5 ± 1.5 kg/m²). All subjects were free from lower limb injury at the time of data collection. All subjects were informed of the possible risks of the investigation which was approved by the Hong Kong Polytechnic University’s Review Board for use of human subjects in research. To maintain consistency with the modeling and experimental evaluation metrics, and validate the FEM by contact pressure measurement, five items of custom-made compression sports device with the same compression design of that in the FEM were involved in this experiment.

Contact pressure data were collected with Flexi Force A201 force sensors (Tekscan Inc., USA). This pressure collection system has been described in our previous studies.¹⁸ Force sensors were adhered to the seven important muscles that flex or extend when running (Figure 2a): the vastus lateralis (VL), vastus medialis (VM), rectus femoris (RF), tibialis anterior (TA), semimembranosus (SE), gastrocnemius lateralis (GL), and gastrocnemius medialis (GM), which is similar to our previous studies.¹⁸

Figure 2.

(a) The locations of the selected muscle points. (b) method of experimental measurement.

Contact pressures were recorded under the condition of bare lower limb at standing position, and wearing sports device at standing, 30°, 60°, 90°, and 120° knee flexion positions. Participants were instructed to put their forefoot on the stacked printing paper packages or a plastic stool, which could easily adjust the height to reach the designed knee flexion angles. This position was held for 8 to 10 seconds for pressure recording, which was taken at each position and repeated seven times. Participants were instructed to hold the support when needed, straighten up the torso, relax their measured leg, and keep their breathing steady during recording. The mean values and standard deviation were given and represented graphically using SPSS17.

Statistics

In the present work, median values and interquartile ranges (IQRs) are presented using boxplot graphs. The Kruskal-Wallis test (KW test) was conducted on all knee flexion angle groups for the specific muscle points. If the KW test was significant, then multiple comparisons were performed among individual pairs of flexion angles with the Nemenyi test, a Tukey type non-parametric test.³⁰ The Pearson test and the Spearman non-parametric test were performed for quantifying correlations of experimental measurements and simulation. The descriptive and correlation statistical tests were run under SPSS 23 (IBM Corporation). The graphs were generated by SPSS 23 and Excel 2016.

Output of the simulation

Within the aim of characterizing the compression pressure distribution from the simulation with the experimental validation, two sets of analysis data from the simulation output were presented:

the mean pressure on the seven important muscle points during knee flexion;

the mean pressure P, and pressure gradient G at the commonly used measuring points along the anterior, medial, posterior, and lateral parts of the lower limb, including B (ankle), B1 (area at which the Achilles tendon changes into the calf muscles), C (calf at its maximum girth), D (just below the tibial tuberosity), LT (lower thigh), MT (mid thigh), and UT (upper thigh). The mean pressure P (anterior side: $P_{i, A}^{θ}$ , medial side: $P_{i, M}^{θ}$ , posterior side: $P_{i, P}^{θ}$ , lateral side: $P_{i, L}^{θ}$ ) was obtained by the mean pressure over an area with radius of 2.5 cm at the measurement point. The cross-section pressure gradient G (anterior side: $G_{i, A}^{θ}$ , medial side: $G_{i, M}^{θ}$ , posterior side: $G_{i, P}^{θ}$ , lateral side: $G_{i, L}^{θ}$ ) was the rate of change of the mean pressure P of the four-side values on the selected cross-section. In this study, the cross-section region of the ankle (B) is selected as the baseline section, as shown in Equation (1). The cross-section pressure gradient deviation ΔG on the selected side regions, as shown in Equation (2).

G_{i, j}^{θ} = \frac{P_{i, j}^{θ}}{(P_{B, A}^{θ} + P_{B, M}^{θ} + P_{B, P}^{θ} + P_{B, L}^{θ})}

(1)

Δ G_{i, s}^{θ} = G_{i, s}^{θ} - G_{i, j}^{θ} = \frac{P_{i, s}^{θ} - P_{i, j}^{θ}}{(P_{B, A}^{θ} + P_{B, M}^{θ} + P_{B, P}^{θ} + P_{B, L}^{θ})}

(2)

\begin{matrix} θ = 0^{\circ}, 30^{\circ}, 60^{\circ}, 90^{\circ}, 120^{\circ}, \\ i = B, B 1, C, D, LT, MT, UT, \\ j = A, M, P, L \end{matrix}

where θ is the angle of knee flexion, i is the height level of the measuring points along the lower limb, and j is the four sides of the lower limb.

Results

A 3D lower limb device FEM was introduced to evaluate the garment–skin interface pressure during deep knee flexion. The nonlinear response induced by external compression of the lower limb was illustrated and validated.

Experimental contact pressure distribution and validation

Figure 3 depicts the local pressure values predicted by the FEM with the sports device on the important muscle points from standing to deep knee flexion with the measurement data. When the knee flexed at 30°, the GM (KW test: p ≤ 0.01) and SE (p ≤ 0.001) is at a significantly higher pressure (GM 107% and SE 530%) than in standing position (Figure 3(b) and 3(f)), which are on the back of the calf and thigh. With the knee flexed at 60°, contact pressure was only partially but significantly decreased to 66% (p ≤ 0.01 compared with 30° angle point measurements) at GM on the calf. The average contact pressure of GL on the calf was gradually and significantly elevated from 1032 Pa at 60°, to 1620 Pa at 90° (p ≤ 0.01 compared with 60°), and peaked at 1864 Pa at 120° (p ≤ 0.001 compared with 60°). A generally minor increase in contact pressure was observed for the TA and RF positions, which are both on the front of the calf and the thigh, from 30° during knee flexion experiments (Figure 3(a) and 3(g)). For the VM and VL, no significant changes during knee flexion testing were observed. The simulation results for the TA are a little lower than the experimental data, while the others are within the range of the experimental data. For the contact pressure among the seven muscles, the pressure distribution on the calf muscles, GM and GL, was the most highly influenced by deep knee flexion in the experimental tests.

Figure 3.

Contact pressure of FEM simulation and experiment in standing and at different knee flexion angle positions (30°, 60°, 90°, and 120°) on the muscle points: (a) tibialis anterior (TA), (b) gastrocnemius medialis (GM), (c) gastrocnemius lateralis (GL), (d) vastus medialis (VM), (e) vastus lateralis (VL), (f) semimembranosus (SE), and (g) rectus femoris (RF). KW Nemenyi test: *p ≤ 0.01, **p ≤ 0.001.

There was a significant Pearson correlation (r = 0.603, N = 35, p < 0.001) and Spearman rank correlation (r = 0674, N = 35, p < 0.001) between the median pressure and the FEM predicted pressure on the lower limb (Figure 4). The predicted pressure distribution generally followed the similar pattern as the experimental ones.

Figure 4.

FEM predicted pressure is positively correlated with the experimental median pressure on the selected muscle points of the lower limb (Pearson r = 0.603, p < 0.001; Spearman r = 0674, p < 0.001).

Predicted contact pressure distribution during deep knee flexion

General distributions of simulated contact pressure on the lower limb for different knee flexion angles are depicted in Figure 5. To facilitate the comparison of observation of the lower limb with different knee flexion angles, the simulated pressure distributions of the lower limb in Figure 5 are presented in undeformed form, that is, in standing form, using FEM software. On the knee region, an area of high pressure was observed on the patella during knee flexion, and a correspondingly large area of high pressure developed on the popliteal area at the 120^o knee angle. For the anterior of the leg, these figures show that the anterior crest of the tibia generated a narrow band area with high contact pressure, whereas its internal surface generated a parallel narrow band area with low pressure. For the posterior of the leg, high pressure areas were observed on the prominent parts of GM and GL.

Figure 5.

Distributions of predicted contact pressure on the skin surface for different flexion angles of the knee in undeformed form using FEM.

Figure 6.

(a) Predicted average interface pressure profile of compression sports device along anterior, medial, posterior, and lateral sides at landmarks in standing posture. Predicted average pressure gradients of compression sports device worn in standing and knee flexion postures of 30°, 60°, 90°, and 120° angles at landmarks: (b) cross-section pressure gradient deviation on landmarks along anterior and posterior sides, and (c) lateral and medial sides.

Contact pressure distribution and variation distributions were analyzed along four different side and different cross-section regions. Figure 5(a) demonstrates the contact pressure distribution on the anterior, medial, posterior and lateral sides indicating the resultant compression of the different locations (B, B1, C, D, LT, MT, UT) acting on the lower limb. At standing position, the range of contact pressure (mean±SD) along the four directions from ankle to the thigh as shown in Table 1 and Figure 5(a). For the ankle interface pressure as shown in Figure 5(a), the average pressure of the sports device around the ankle was 944 Pa. The predicted lowest pressure was found at the medial site 661 Pa (70% of the four-side average pressure at ankle), and the highest at the pretibial area (1265 Pa, 134%).

Table 1.

Pressures (mean±SD) of the landmarks along four sides of the lower limb

	B	B1	C	D	LT	MT	UT
Average pressure (Pa)	944	564	685	316	391	341	319
±SD (Pa)	323	93	141	107	65	54	150

The variation in contact pressure distributions on the selected locations at medial view and anterior view are shown on Figure 5(b) and 5(c), respectively. For the average pressure and cross-section pressure gradient deviation on the lower limb, comparing the pressure deviation at the landmarks of four sides along different height levels in Figure 5(a) revealed that the pressure varied more on the ankle (B) and calf (B1, C, and D) regions than on the thigh regions (LT, MT, and UT) within the range of ±15% during deep knee flexion. The contact pressure deviation ΔG dramatically increased on the anterior side of B1 (96%) and D (115%) at 60^o, then dropped to 21% and −6%, respectively, while on their neighboring regions B and C, the cross-section pressure gradient deviation ΔG reached the lowest values at 120^o knee flexion, at the anterior direction (36%) and posterior direction (16%), respectively. The ΔG generally increased from standing position (baseline: 0%) to the 120^o knee flexion angle (28%) at the medial side of the ankle (B), in Figure 5(c).

Discussion

The purpose of this study was to evaluate how the deep knee flexion behavior results in lightweight compression effects on the lower limb wearing a compression device. To contribute to solving these problems, the present study was designed not only to measure directly how contact pressure on the selected important muscle points is distributed, but also to illustrate the whole pressure distribution exerted by the compression device during knee flexion. Experimental pressure testing and FE simulation were applied to quantify the contact pressure and cross-section pressure gradient deviation to improve the understanding of the effect of lightweight compression devices on the lower limb during knee flexion. This has several advantages for original design evaluation in enabling the examination of dynamic biomechanical modeling, which is difficult to examine using direct experimental measurement. This proposed approach has the potential to investigate fully the dynamic effects on calf muscle pumping with high performance sportswear or devices, injury prevention with knee protectors, and reducing the metabolic rate of elderly people with walking assistive devices.

Although interface pressure is usually used as the most important characteristic of compression garments, the changes of pressure distributions on the whole leg during knee flexion remain largely unexplored. The present study is the first to define the cross-section pressure gradient deviation to quantify the side deviation from the mean of four-side compression on the selected cross-section compared with the four-side values at the ankle part as the common denominator. This has practical implications for evaluating the change of pressure distribution on the whole leg at different knee flexion angles. The novel approach of the cross-section pressure gradient deviation enabled quantitative evaluations of the dynamic compression effects, by using deviation modalities that demonstrate the change of compression effect by the load along four axial pathways and cross-section direction on the lower limb. Hence, the compression effects that are affected by deep knee flexion as well as the irregular body surface could be estimated. The findings in the current study can facilitate the dynamic ergonomic design of lower limb devices with different purpose.

In general, our simulated results show a gradual decrease in pressure distribution at standing position of four axial pathways along the lower limb, and high compression pressure on the C point over the calf, as shown in Figure 4(a) and Table 1. The compression device generally presented a gradual pressure profile with a high pressure at the ankle (944 Pa), median pressure at the calf (685 Pa) and a low pressure at the thigh (319 Pa); while a slightly higher pressure at the calf (685 Pa) compared with the adjacent B1 (564 Pa) and D (316 Pa) positions. This designed pressure profile may probably integrate the advantages of graduated compression from the ankle to the thigh to improve the circulation, and negative gradient compression with stronger pressure over the calf to enhance the calf muscle pumping. The simulated results of pressure distribution have a good agreement with our previous study¹⁸ as well as other published research.^11,12 Meanwhile, the observed average contact pressure of calf muscle points (TA, GM, GL) is generally higher than those of thigh muscle points (VM, VL, SE, RF), as shown in Figure 3.

Comparing the predicted series at B1, C, and D on the calf part, there are different patterns on the front and the back during knee flexion. At a 60^o angle during knee flexion, the pressure deviation achieved the highest point (B1 and D) at the anterior direction of the calf, which resembled an M shape, as shown in Figure 4(b). One possible reason for the relatively large variation of cross-section pressure gradient deviation at the anterior of calf, which resembled an M shape, may well be the large irregular clothing deformation caused by the negative gradient design of the lower limb device that exerted high compression on the calf and relative low compression on its neighboring regions. In addition, the stretch of the clothing during knee flexion, the large local curvature, and the bony structures underneath the anterior calf may be the reason why the anterior part reached the high pressure and large cross-section pressure gradient deviation compared with other directions on the same cross section. Consequently, the sports devices based on graduated compression on the whole leg with high compression design on the calf may generate large variation of interface pressure at the anterior part of the calf during knee flexion, which has large local curvature and sustaining bony structures.

The study described here has several limitations that deserve consideration. First, the study used a simplified model without ligaments and tendons, which does not fully represent the contact and movement behavior on the natural knee part. It is therefore suggested that further improvement of the knee flexion modeling should be conducted to advance our understanding of the compression effects on the lower limb during knee flexion behavior. Second, although the simulation results reported a high level of pressure variation on the calf during knee flexion, caused by the high compression design over the calf, evaluating the real benefit to walking or running performance of the current lower limb devices remains unclear and challenging. Furthermore, combining FEM with different fabric properties, structures, and pattern designs for the original design may facilitate the design process of the compression lower limb devices, such as high performance sports tights, knee protection, and walking assistance devices.

Conclusion

An FEM of the male leg with detailed anatomical bony structures and sports lower limb device, was developed to predict interface contact pressure and the cross-section pressure gradient deviation distribution. Nonlinear elastic models of soft tissue, skin, and clothing material were applied in the FEM. The predicted contact pressure agrees well with the original design pressure profile, that is, generally gradual decrease from the ankle to the thigh, and a relatively high level of pressure on the calf. The cross-section pressure gradient deviation was introduced to calculate the pressure deviation from the average four-points pressure on the same cross section and compared with the ankle cross section as the baseline. The 3D FEM and experiments demonstrated that the pressure varied dramatically on the calf part during deep knee flexion. For the ankle part, cross-section pressure gradient deviation generally increased at the medial direction and decreased at the anterior direction from standing to the 120^o angle. The distribution of contact pressure on the thigh part and the lateral side along the whole lower extremities only suffered small changes in amplitude both in the simulation and experiment. The trend of predicted cross-section pressure gradient deviation showed an M shape on the anterior part of the calf, that is, the local pressure gradient deviation on the cross section of the calf had a lower value compared with its neighboring (B1 and D) regions, which were influenced by the higher local pressure design of the device on the calf part, large stretch of clothing of the garments or devices, and the irregular anatomical structure of the leg. The FEM successfully predicted the pressure profile of the original design sports lower limb device, and should further improve the understanding of the interface of body and device and facilitate the engineering process of compression garments and devices. This original design is suitable for several practical applications, such as high performance sportswear and devices for elite athletes, and passive or active walking assistive devices for elderly people.

Footnotes

Acknowledgements

We would like to thank the University of Manchester, Xian Polytechnic University, Hong Kong Polytechnic University, and Hong Kong Research Institute of Textile and Apparel for continuous support.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Key Research and Development Program of China (2018YFC2000903).

ORCID iD

Yinglei Lin

References

Kemmler

von Stengel

Köckritz

, et al. Effect of compression stockings on running performance in men runners. J Strength Cond Res 2009; 23(1): 101–105.

Ali

Caine

Snow

. Graduated compression stockings: physiological and perceptual responses during and after exercise. J Sport Sci 2007; 25(4): 413–419.

Moehrle

Kemmler

Rauschenbach

, et al. Acute and long term effect of compression stockings in athletes with venous insufficiency. Phlebologie 2007; 36(6): 313–319.

Lawrence

Kakkar

. Graduated, static, external compression of the lower-limb – physiological assessment. Brit J Surg 1980; 67(2): 119–121.

Mosti

Partsch

. High compression pressure over the calf is more effective than graduated compression in enhancing venous pump function. Eur J Vasc Endovasc 2012; 44(3): 332–336.

De Glanville

Hamlin

. Positive effect of lower body compression garments on subsequent 40-km cycling time trial performance. J Strength Cond Res 2012; 26(2): 480–486.

Reich-Schupke

Surhoff

Stucker

. Pressure profiles of sport compression stockings. Journal der Deutschen Dermatologischen Gesellschaft = Journal of the German Society of Dermatology : JDDG 2016; 14(5): 495–506.

McLaren

Helmer

RJN

Horne

, et al. Preliminary development of a wearable device for dynamic pressure measurement in garments. Procedia Engineering 2010; 2(2): 3041–3046.

Brophy-Williams

Driller

Shing

, et al. Confounding compression: the effects of posture, sizing and garment type on measured interface pressure in sports compression clothing. J Sport Sci 2015; 33(13): 1403–1410.

10.

Wertheim

Melhuish

Williams

, et al. Movement-related variation in forces under compression stockings. European Journal of Vascular and Endovascular Surgery 1999; 17(4): 334–7.

11.

Wildin

Hui

Esler

, et al. In vivo pressure profiles of thigh-length graduated compression stockings. Brit J Surg 1998; 85(9): 1228–1231.

12.

Liu

Kwok

, et al. Skin pressure profiles and variations with body postural changes beneath medical elastic compression stockings. Int J Dermatol 2007; 46(5): 514–523.

13.

MacRae BA, Laing RM, and Partsch H. General considerations for compression garments in sports: applied pressures and body coverage. In: Compression garments in sports: athletic performance and recovery. Switzerland: Springer, 2016, pp. 1–32.

14.

Partsch H and Jünger M. Evidence for the use of compression hosiery in lymphoedema. In: Template for practice: Compression hosiery in lymphoedema. London: MEP Ltd, 2006, pp. 5–9.

15.

Stolk

van der-Franken

Neumann

HAM

. A method for measuring the dynamic behavior of medical compression hosiery during walking. Dermatol Surg 2004; 30(5): 729–736.

16.

Hirai

Niimi

Iwata

, et al. A comparison of interface pressure and stiffness between elastic stockings and bandages. Phlebology 2009; 24(3): 120–124.

17.

Karakashian

Pike

van Loon

. Computational investigation of the Laplace law in compression therapy. J Biomech 2019; 85: 6–17.

18.

Lin

Choi

Luximon

, et al. Finite element modeling of male leg and sportswear: contact pressure and clothing deformation. Text Res J 2011; 81(14): 1470–1476.

19.

Lin

Luximon

Choi

, et al.A three-dimensional biomechanical model of leg for numerical simulation of skin pressure of sportswear. In: Li

Chen

(eds). Textile bioengineering and informatics symposium proceedings, Vols 1 and 2, Kowloon, Hong Kong: Hong Kong Polytechnic University, 2009, pp. 272–275.

20.

Frauziols

Molimard

Navarro

, et al. Prediction of the biomechanical effects of compression therapy by finite element modeling and ultrasound elastography. IEEE Transactions on Bio-medical Engineering 2015; 62(4): 1011–1019.

21.

Lin

Choi

Zhang

, et al. An optimized design of compression sportswear fabric using numerical simulation and the response surface method. Text Res J 2012; 82(2): 108–116.

22.

Liu

Kwok

, et al. A three-dimensional biomechanical model for numerical simulation of dynamic pressure functional performances of graduated compression stocking (GCS). Fiber Polym 2006; 7(4): 389–397.

23.

Dai

X-Q

Zhang

, et al. Effect of sock on biomechanical responses of foot during walking. Clinical Biomechanics 2006; 21(3): 314–321.

24.

Lin

. Computational modeling and mechanical analysis of compression sportswear, Kowloon, Hong Kong: The Hong Kong Polytechnic University, 2011.

25.

Bosboom

EMH

Hesselink

MKC

Oomens

CWJ

, et al. Passive transverse mechanical properties of skeletal muscle under in vivo compression. J Biomech 2001; 34(10): 1365–1368.

26.

Ogden

Elastic deformations in rubberlike solids. In: Hopkins

Sewell

(eds). Mechanics of solids, The Rodney Hill 60th anniversary volume, Oxford: Pergamon Press, 1982, pp. 499–537.

27.

Barbarino

Jabareen

Mazza

. Experimental and numerical study on the mechanical behavior of the superficial layers of the face. Skin Res Technol 2011; 17(4): 434–444.

28.

Victor

Labey

Wong

, et al. The influence of muscle load on tibiofemoral knee kinematics. J Orthop Res 2010; 28(4): 419–428.

29.

Belvedere

Leardini

Ensini

, et al. Three-dimensional patellar motion at the natural knee during passive flexion/extension. An in vitro study. J Orthop Res 2009; 27(11): 1426–1431.

30.

Zar

. Biostatistical analysis, London: Prentice-Hall International, 1996.