Previous studies indicate that flexible footwear, which mimics the biomechanics of walking barefoot, results in decreased knee loads in patients with knee osteoarthritis (OA) during walking. However, the effect of flexible footwear on other activities of daily living, such as descending stairs, remains unclear. Our objective was to evaluate the influence of inexpensive and minimalist footwear (Moleca) on knee adduction moment (KAM) during stair descent of elderly women with and without knee OA.
Thirty-four elderly women were equally divided into an OA group and a control group (CG). Stair descent was evaluated in barefoot condition, while wearing the Moleca, and while wearing heeled shoes. Kinematics and ground reaction forces were measured to calculate KAM by using inverse dynamics.
The OA group experienced a higher KAM during midstance under the barefoot condition (233.3%; P = 0.028), the Moleca (379.2%; P = 0.004), and heeled shoes (217.6%; P = 0.007). The OA group had a similar knee load during early, mid, and late stance with the Moleca compared with the barefoot condition. Heeled shoes increased the knee loads during the early-stance (versus barefoot [16.7%; P < 0.001] and versus the Moleca [15.5%; P < 0.001]), midstance (versus barefoot [8.6%; P = 0.014] and versus the Moleca [9.5%; P = 0.010]), and late-stance phase (versus barefoot [10.6%; P = 0.003] and versus the Moleca [9.2%; P < 0.001]). In the CG, the Moleca produced a knee load similar to the barefoot condition only during the early-stance phase.
Besides the general foot protection, the inexpensive and minimalist footwear contributes to decreasing knee loads in elderly women with OA during stair descent. The loads are similar to the barefoot condition and effectively decreased when compared with heeled shoes.
Among the various types of arthritis, osteoarthritis (OA) is the most common and the major cause of chronic musculoskeletal pain and disability in the elderly population worldwide (1). The knee is the most affected joint in the lower extremities (2), particularly the medial compartment, which is greatly responsible for load attenuation during locomotion (3).
In daily living activities, patients with knee OA often have to change locomotion direction and speed, manage slopes, and ascend and descend steps at home, on buses, or in public and private places with irregular pavements. Although these activities could be painful or avoided due to patients' functional deficit, such daily locomotor tasks play an important role in the functionality and independence of the elderly, and their ability to perform routine daily tasks is an important factor for good quality of life (4). OA interferes directly with the individual's ability to perform these functional activities and tasks expected of an independent adult (5).From a mechanical point of view, the motor task of stair descent requires a coordinated action of the ankle and knee joints, involving a higher vertical acceleration of the body (6, 7), increased mechanical demand (7), and higher joint forces and moments (6). Therefore, it is suggested that the potential harm to elderly people with knee OA during this task can be even greater than during level walking.
Reduction in knee loading is considered one of the most important therapeutic objectives in treating OA (8), as overload increases the risk of incidence and progression of the disease (9, 10). In addition, the flexibility of barefoot locomotion has been shown to reduce the external knee adduction moment (KAM), which is a biomechanical parameter that often increases in individuals with knee OA (8–10). However, the opportunity to perform daily activities such as stair descent while barefoot is rare. In addition, modern shoes, usually heel elevated with a rigid sole, are commonly used by the majority of women in their everyday lives (11). There is evidence that these types of shoes generate higher knee torques during locomotion when compared to barefoot locomotion (12), and because of their lower malleability, modern shoes can damage knee joints with OA, as they do not reproduce the flexibility of the movements of the foot when it is bare or that of the lower extremity as a whole (8).
Based on the conclusions of studies regarding the biomechanical consequences of modern shoes and of barefoot locomotion on knee mechanics, it is reasonable to suggest that the use of a shoe that can faithfully reproduce the mechanical properties of barefoot conditions may minimize the damage caused by excessive intraarticular knee stresses in individuals with knee OA, especially during tasks that demand higher mechanical torques, such as stair descent.
Therefore, the purpose of the present study was to evaluate and compare the influence of a modern heeled shoe with the barefoot condition and inexpensive, flexible, nonheeled footwear (Moleca), already produced and used on a large scale, on KAM during stair descent in elderly women with and without knee OA. The initial hypotheses were: 1) the Moleca would result in external KAM similar to the barefoot condition in stair descent; 2) modern heeled shoes, with a more rigid sole, would increase the magnitude of external KAM; and 3) the acute effects of the Moleca in elderly women with knee OA would promote external KAM similar to those without OA.
Significance & Innovations
It has already been shown that the mechanical loads (knee adduction moment) in the knee of elderly people with osteoarthritis (OA) are decreased during barefoot locomotion; however, elderly people do not perform their daily living activities barefoot.
We have proposed the use of an inexpensive (US $9) and minimalist footwear (flexible and nonheeled) to mimic barefoot locomotion, and thus reduce the adduction moment that is related to the progression of knee OA.
This minimalist footwear proved to be efficient to acutely reduce the knee loads during stair descent in both elderly women with and without knee OA compared to barefoot and modern heeled shoes.
MATERIALS AND METHODS
A sample power calculation was carried out a priori based on the primary outcome (KAM), considering an F test statistical design for repeated measures (between and within effects) with a moderate size effect of 0.25, a power of 80%, and an alpha error of 5%, resulting in a total sample size of 34 subjects, divided equally in 2 groups (17 subjects in each group).
Thirty-four elderly women were equally divided into 2 groups: an OA group and a control group (CG). The OA group was comprised of women (mean ± SD age 65 ± 6 years, mean ± SD body mass 70.9 ± 8.0 kg, mean ± SD height 1.56 ± 0.05 meters, and mean ± SD body mass index [BMI] 29.2 ± 3.3 kg/m2) with a diagnosis of OA on the medial compartment of the knee, graded 2 (48%) and 3 (52%) according to the Kellgren/Lawrence (K/L) scale (13). The CG included elderly women without diagnosed knee OA (mean ± SD age 66 ± 4 years, mean ± SD body mass 64.2 ± 9.0 kg, mean ± SD height 1.57 ± 0.07 meters, and mean ± SD BMI 26.1 ± 3.0 kg/m2). The groups were not statistically different in terms of age (P = 0.669) or height (P = 0.640). Although the OA group had a larger body mass (P = 0.029) and BMI (P = 0.006), the participants of both groups were classified as overweight. All of the participants gave their written informed consent approved by the local ethical review committee (765/07).
The American College of Rheumatology criteria (14) were used to select the women in the OA group. All of the women in this group had symptoms of knee pain during the week prior to the evaluation. All of the individuals from both the OA group and the CG had bilateral, weight-bearing, full-extension, anteroposterior knee radiographs, which were classified according to the K/L scale (13). To better characterize the knee function of the participants, the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaire (15) was applied. The mean ± SD WOMAC scores relating to the 72 hours prior to the biomechanical evaluation of both groups are shown in Table 1.
Table 1. Descriptive statistics of the Western Ontario and McMaster Universities Osteoarthritis Index scores for the OA and control groups*
Major exclusion criteria for both groups were a BMI greater than 35 kg/m2 (8, 16, 17); the presence of knee instability; hip or ankle OA; disabling or systemic arthritis; neuromuscular dysfunctions (18); rheumatoid arthritis (19); inability to provide consistent information; use of prostheses and/or orthoses in the lower extremities (20); muscular injuries or a history of surgery in the knees, ankles, or hips (19); the use of any assistive walking devices such as a walking stick or crutches for independent gait; presence of limping; and inability to go up or down stairs (21).
All of the individuals were asked to bring one pair of closed footwear with a rounded shape in the front, a high heel between 45 and 50 mm, and a wide-base heel approximately 40 mm × 40 mm, which were habitually used by them for at least 4 hours 3 times a week.
Stair descent analysis protocol.
Passive–reflective markers were placed on the subject using a standard Cleveland Clinic marker set (anterior superior iliac spine, superior aspect of the greater trochanter, lateral knee joint line, lateral malleolus, calcaneus, and head of the fifth metatarsal joint) (22). Extra markers were placed bilaterally at the medial knee joint line, medial malleolus, and first metatarsal joint for the static standing trial in order to determine relative joint centers of rotation for the knee, ankle, and longitudinal axis of the foot. These extra markers were removed in the gait trial. In addition, 3 noncolinear reflective markers were fixed at 2 squares, forming sets of technique cluster. One of these was placed in the lateral thigh and the other over the shank. The laboratory coordinate system was established at one corner of the force plate, and all of the initial calculations were based on this coordinate system. Based on the surface markers, each lower extremity segment (foot, shank, and thigh) was modeled as a rigid body with a local coordinate system that coincided with anatomic axes, and translations and rotations of each segment were reported relative to neutral positions defined during the initial standing static trial.
Three-dimensional (3-D) marker displacements were evaluated with 6 infrared cameras (Optitrack FLEX: V100). The automatic digitizing process, the 3-D reconstruction of the markers' positions, and the filtering of kinematic data were performed using Arena software. The accuracy of the full-volume test for the Optitrack System was performed, as suggested by Chiari et al (2005) (23), and the results showed errors of less than 5.6 mm, similar to the majority of the commercial systems usually used in biomechanics (24). Ground reaction forces were acquired by a force plate (AMTI OR-6 – 1000) embedded in the floor at the end of the last stair step. Data acquisition was synchronized and sampled by an A/D card (AMTI, DT 3002, 12 bits) at 100 Hz.
Before data acquisition, all of the participants received the same instructions: to descend a staircase of 5 steps without using the handrail, beginning the task with the opposite extremity to be evaluated, and positioning one foot on each step during the descent. The subjects were instructed to descend the staircase as they were used to doing on a daily basis at their self-selected cadence. Each step of the staircase was 60 cm wide, 32 cm deep (run), and 20 cm high (rise), with a slope of 32 degrees between the rise and run. Although all of the individuals were able to perform the stair descent trials independently, we allowed them to use the handrail if they felt unsecure (25), but excluded this particular trial from the statistical analysis. There was no significant difference between the groups in the descent cadences during the trials (CG: mean ± SD 99 ± 7 steps/minute, OA group: mean ± SD 85 ± 11 steps/minute; P = 0.061). The last step performed by the selected extremity was analyzed. In the OA group, the lower extremity of the index knee performed the last step to the force plate, and in the CG, the lower extremity that performed the last step was chosen randomly by a simple drawing using slips of paper. For each participant, 5 valid trials per condition were recorded after total adaptation to the laboratory environment, the shoes (26), and the staircase (21), and after performing a consistent locomotion pattern. The mean of the 5 trials was used for statistical purposes.
The order of the stair descent conditions was randomly established for the barefoot, Moleca, and modern heeled shoes. The Moleca (Calçados Beira Rio) footwear is a women's minimalist flat walking shoe without a heel, with a thin 5-mm antislip flexible rubber sole and a 3-mm internal wedge of ethylene vinyl acetate, with an upper body (heel cap, quarter, vamp, and toe cap) made out of double canvas (Figure 1). Its mean ± SD weight is 0.172 ± 0.019 kg, ranging from 0.142 to 0.193 kg depending on the size. The Moleca shoes used in this study are property of the present laboratory. The modern heeled shoes, which are habitually used by women, had the following characteristics: an upper body made of soft leather, a hard leather outsole (low flexibility) with a wide-base heel (mean ± SD width × length: 4.08 ± 0.76 mm × 4.46 ± 0.69 mm for the OA group and 4.13 ± 0.83 mm × 4.75 ± 0.78 mm for the CG; width: P = 0.343, length: P = 0.899) (12, 27–29), and a mean ± SD heel height of 40.0 ± 9.0 mm for the OA group and 40.4 ± 9.0 mm for the CG (P = 0.102).
Stair descent data processing and statistics.
Kinematic data were processed using a second-order low-pass filter with a cutoff frequency of 6 Hz. The ground reaction force data were processed using a zero-lag low-pass Butterworth fourth-order filter with a cutoff frequency of 20 Hz. The inverse dynamics approach was used to calculate the external knee moment in the frontal plane, using Visual3D software (C-motion). The inertial properties were based on Dempster's standard regression equations (30). The moment of inertia and location of the center of mass were computed assuming the thigh and shank segments as a cylindrical geometric shape. The knee was examined as a bicondylar-type joint and, according to Perry (1992) (31), adopted with 3 planes of motion: sagittal, transverse, and coronal.
We calculated KAM variables in a custom-written math function in Matlab (MathWorks): the first KAM peak at the weight acceptance phase (∼20% of stance phase), the KAM at the forward continuance phase (corresponding to horizontal body displacement of ∼20% to 55% of stance phase), the second KAM peak at the propulsion phase (∼80% of stance phase), and the non-normalized adduction angular impulse (32). Data of only one lower extremity per subject were analyzed and compared. For the CG, the lower extremity was randomly chosen; for the OA group, the selected extremity corresponded to the symptomatic knee in subjects with unilateral OA and to the most symptomatic knee in subjects with bilateral OA (8, 16, 17).
Statistical tests included a two-way analysis of variance (2 [groups] × 3 [gait conditions]) to compare KAM variables, followed by the Newman-Keuls post hoc test (α = 5%). Anthropometric and demographic characteristics were compared between the groups using t-test, and WOMAC scores were compared using the Mann-Whitney test (α = 5%; Statistica, version 8, Statsoft).
Table 2 and Figure 2 show that the first and second KAM peaks were not statistically different between the groups; however, the KAM magnitude during the forward continuance phase was higher in the OA group for all conditions, suggesting a more significant articular load, even when barefoot or with the flexible shoes. In the forward continuance phase, the KAM of the OA group was 1.3 times greater (233.3%) in the barefoot condition (P = 0.028), 2.8 times greater (379.2%) with the Moleca (P = 0.004), and 1.1 times greater (217.6%) while wearing modern shoes (P = 0.007) than that of the CG.
Table 2. Descriptive statistics of the KAM variables and comparisons between the OA group and the CG for the conditions of barefoot, the Moleca, and heeled shoes*
During the weight acceptance phase, in both groups the Moleca condition resulted in a first KAM peak that was not different from the barefoot condition (CG: P = 0.132, OA group: P = 0.487). In the CG, as expected, heeled shoes increased this variable by 15.2% when compared to the barefoot condition and by 22.2% when compared to the Moleca condition (P < 0.001). Similarly, in the OA group there was an increase of 18.4% in the first KAM peak when wearing heeled shoes compared to the barefoot condition (P < 0.001), and an increase of 21.4% in the first KAM peak when compared to the Moleca footwear (P < 0.001).
During the forward continuance phase, the KAM in the CG was 39.1% lower with the Moleca compared to the barefoot condition (P = 0.016) and 48.1% lower compared to heeled shoes (P = 0.002), while in the OA group this variable was 10.4% lower with the Moleca than with heeled shoes (P < 0.001) and heeled shoes were not different from barefoot.
During the propulsion phase, the second KAM peak in the CG was 18.3% lower with the Moleca than both the barefoot (P = 0.024) and heeled (P = 0.010) conditions, while in the OA group this variable was 9.2% lower with the Moleca than with heeled shoes (P < 0.001) and was not different from barefoot.
The non-normalized knee adduction impulse in the CG was 12.4% lower while barefoot when compared to heeled shoes (P = 0.033), whereas the Moleca footwear decreased this variable by 16.2% in comparison to the barefoot condition (P = 0.015), and was 26.6% lower when compared to heeled shoes (P < 0.001). In the OA group, the impulse was 15.2% lower while barefoot (P < 0.001) and 16.3% lower with the Moleca footwear (P < 0.001) when both were compared to the heeled shoes.
The purpose of the present study was to evaluate and compare the influence of inexpensive, flexible, nonheeled footwear (Moleca) with a modern heeled shoe and the barefoot condition on the KAM during stair descent in elderly women with and without knee OA. The main findings supported our initial hypotheses that the Moleca would result in external KAM similar to the barefoot condition in stair descent and that modern heeled shoes, with a more rigid sole, would increase the magnitude of external KAM. Wearing the Moleca resulted in a similar knee load during stair descent as being barefoot in women with OA, whereas with heeled shoes there was an increase in knee load compared to both the Moleca and being barefoot. Although the effects of the Moleca on KAM magnitude were similar in elderly women with and without OA during the weight acceptance and propulsion phases, in the forward continuance phase the Moleca decreased the KAM magnitude compared to the barefoot and heeled shoe conditions. In the OA group, the Moleca behaved in the same manner as in the barefoot condition.
In all stair descent conditions, the OA group presented a worse load absorption capacity in the knee joint during the forward continuance, where there are high load magnitudes for approximately 40% of the stance phase (between 30% and 70% of the stance phase) (33). This result can be explained by the poor proprioceptive acuity related to the presence of knee pain and its severity and by the physical functional deficits in patients with knee OA (34). Therefore, this can also be attributed to the subject's attempt (without success) to attenuate the vertical acceleration during stance phase, reducing the stress in the injured joint. Chen et al (2003) (35) also have found similar results, i.e., worse weight acceptance during the midstance phase in elderly women with knee OA.
In addition, according to Chen et al (2003) (35), the high degree of coactivation of antagonist muscles present in patients with knee OA may also play a role in the abnormal joint loading, primarily during midstance. This result can explain the higher moment magnitudes observed in the present study in this particular support phase in the stair descent task.
The increased rigidity of the soles of the heeled shoes (36) and not the heel elevation could be responsible for the higher KAM observed while descending steps compared to the Moleca and barefoot conditions. This sole rigidity, primarily during the weight acceptance and propulsion phases, may increase the ground reaction force magnitude, and consequently the KAM peaks. In particular, the ankle is kept in an extended position during the initial phase of stair descent (the weight acceptance phase). Therefore, the contact of the high heel to the step is minimal, if any, and the forefoot, a more flexible foot area but restricted by the sole stiffness of the heeled shoe, is mainly responsible for the stair contact and is the first to attenuate the loads imposed by the task. This forefoot rigidity may have caused the higher KAM, regardless of the group. In addition, the constantly extended ankle position imposed by the heeled shoe during the stair descent can make the eccentric action of the triceps surae impaired, and consequently the load attenuation ability in the foot–ankle complex could be negatively affected, increasing the knee loads.
The choice of using high-heeled footwear in the present study was due to the fact that this type of footwear is preferred by 70% of women in their everyday lives; they use it for at least 4 hours 3 times a week, and the majority of this modern footwear for women has heels on the posterior part and much more rigid soles (11). We decided not to standardize the brand of high-heeled footwear because using a new standard pair would cause great interference in the biomechanical data (37). Although the shoe characteristics were controlled and similar for all of the subjects, the variability between women's shoes may have influenced the results.
Recently, Shakoor et al (2010) (38) tested flexible sports footwear of a renowned brand that demonstrated the same mechanical advantages observed in the present study by the use of the Moleca for elderly individuals with OA, but during gait. However, it should be pointed out that the sports footwear tested by Shakoor et al (H-street, Puma) (38), despite demonstrating similar joint loads as the barefoot gait and being commercially available, presents a substantially higher cost (US $90), thus being less accessible for the majority of the elderly population.
On the other hand, in addition to the mechanical advantages of the Moleca in generating lower KAM peaks than those generated by the modern heeled shoes and similar to the barefoot condition, this flexible footwear has already been produced on a large scale in Brazil since 1986, is usually worn by a large number of elderly people, and costs approximately US $9. This fact makes the use of these shoes not only viable but, above all, efficient for the reduction of the loading on the knee joint in elderly women both with and without OA.
Trombini-Souza et al (2011) (39) evaluated the same footwear used in this study during gait. As a result, the authors found that this footwear demonstrated similar joint loads as the barefoot gait, whereas on the other hand, the heeled shoes showed increased KAM compared to the Moleca and barefoot condition. While descending stairs occurs infrequently in daily life, this condition reflects an extended single-leg stance task with a strong demand of balance and quadriceps strength, which are impaired in OA patients. It is precisely in these situations of greater mechanical demands that there may be further deterioration of the knee joint cartilage. For this reason, our group decided to conduct the present study and contribute to the understanding of this shoe role in locomotion.
In the literature (40, 41), we find descriptions of the greater benefit of the wider sensory perception allowed by the barefoot condition. This greater amount of sensory information coming from the foot contact with the ground triggers neuromuscular reflexes that help minimize the impact and the proximal joint loads. The Moleca shoe possesses a flat and thin rubber sole that allows not only greater sensory perception to the diverse types of ground, but at the same time providing external protection to the feet. This potential benefit of the Moleca shoe may also explain the decrease in KAM while wearing the Moleca compared to both heeled shoes and barefoot conditions in elderly women with OA. Despite the greater sensory perception benefit during barefoot activities, which may contribute to minimizing lower extremity loads, this condition may cause discomfort and insecurity (perhaps for lack of practice) in elderly people, especially in women with knee OA, and consequently could change the KAM magnitude compared to the Moleca shoes. This former shoe is a closed flexible unit of double canvas footwear that potentially minimizes discomfort, thus leading to a more natural foot rollover mechanism. On the other hand, the high-heeled shoes restrict part of the foot's degrees of movement, which may contribute to the greater knee loading (27, 28).
Although the results of this study showed evidence of a decrease of instantaneous loading (peak external adduction moment) and temporal loading (angular adduction impulse) in the knee, these findings were observed only as acute effects. Based on these promising results, future studies should investigate the chronic therapeutic effects of this flexible low-cost footwear on lower extremity biomechanics, structural integrity of the osteochondral tissue, clinical aspects such as pain and inflammatory recurrences, and functionality in the activities of daily living of patients with knee OA. After the corroboration of the long-term results, the Moleca could be prescribed for an extensive range of treatments, considering the onset and delay of the progression of knee OA. The Moleca offers greater advantages than the majority of other treatments, as it provides noninvasive, low-cost, and comfortable therapeutic benefits.
In summary, in the elderly women in our study with and without knee OA, the Moleca resulted in short-term reduction of KAM compared with typical modern heeled shoes, and both the Moleca and the barefoot conditions generated similar KAM during the forward continuance phase.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Sacco had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Sacco, Trombini-Souza, Butugan, Pássaro.
Acquisition of data. Sacco, Trombini-Souza, Pássaro, Arnone.
Analysis and interpretation of data. Sacco, Trombini-Souza, Butugan, Arnone, Fuller.
The authors are grateful to Aline A. Kimura and Ana Paula Ribeiro for help with data acquisition.