DOI: https://doi.org/10.21203/rs.3.rs-2067200/v1
Background: Knee osteoarthritis (KOA), one of the most common musculoskeletal diseases in older adults, is associated with a high incidence of falls. Similarly, toe grip strength (TGS) is associated with a history of falls in older adults; however, the relationship between TGS and falls in older adults with KOA who are at risk of falling is not known. Therefore, this study aimed to determine if TGS is associated with a history of falls in older adults with KOA.
Methods: The study participants, older adults with KOA scheduled to undergo unilateral total knee arthroplasty (TKA), were divided into two groups: non-fall (n=256) and fall groups (n=74). Descriptive data, fall-related assessments, modified Fall Efficacy Scale (mFES), radiographic data, pain, and physical function including TGS were evaluated. The assessment was conducted on the day before performing TKA. Mann–Whitney and chi-squared tests were performed to compare the two groups. Multiple logistic regression analysis was performed to determine the association of each outcome with the presence or absence of falls.
Results: Mann–Whitney U test revealed statistically significant differences in height, TGS on the affected and unaffected sides, and mFES. Multiple logistic regression analysis revealed that the incidence of fall history is associated with TGS on the affected side.
Conclusions: Our results indicate that increasing age and TGS are related to a history of falls in older adults with KOA. The significance of evaluating TGS among patients with KOA in routine clinical practice was demonstrated.
Falls in older adults can lead to injuries such as bone fractures and thus significant deterioration of physical function. Over 30% of people, aged 65 or older, experience approximately one fall every year [1]. Knee osteoarthritis (KOA) is one of the most common musculoskeletal diseases in older adults. It has been reported that an increased history of falls and hip fractures is associated with increasing knee pain, and the incidence of non-vertebral fractures is 1.6 times higher in older adults with KOA than in those without [2]. The incidence of falls in older adults with KOA is approximately 30% higher than that in healthy older adults, and approximately half of individuals aged 60 years or older fall at least once annually [3, 4]. Older adults with KOA and a fall history are more likely to fall after total knee arthroplasty (TKA) [5, 6]. As older adults with KOA are more likely to fall, interventions to prevent falls are important to maintain a high quality of life.
As mentioned, older adults with KOA have a high incidence of falls; however, related risk factors remain undetermined in this population. Knee pain, impaired balance, muscle weakness including that of the knee extensors, and presence of comorbidities have been reported as risk factors for falls in individuals with KOA [7]. Falls in individuals with end-stage KOA are associated with walking ability [8]. Previous studies on older adults without KOA indicate that clinical benchmarks, such as Timed Up and Go (TUG) [9, 10], fall-related self-efficacy [11], knee extension strength [12], and toe grip strength (TGS) [13–15] are associated with the risk of a fall. However, these outcomes may be even stronger predictors in individuals with KOA, who generally have impaired mobility and muscle strength.
Among the aforementioned risk factors for falls, particular attention has been paid to TGS. Many studies reported an association between TGS and falls in older adults. With increasing age, TGS in older adults deteriorates [16–18], resulting in decreased walking speed and static balance ability [19–21]. Toe grip strength training in nursing home residents led to significant improvements in the fall risk index. A significant increase in TGS was noted in the intervention group compared to that in the non-intervention group [22]. Therefore, it is clinically important to evaluate and strengthen TGS to prevent falls in older adults.
To date, no study has investigated the relationship between falls and TGS in older adults with KOA who are at risk of falling owing to lower limb muscle weakness and impaired balance. We hypothesized that the history and frequency of falls in older adults with KOA would be associated with TGS. Therefore, this study aimed to determine whether TGS is associated with a history of falls in older adults with KOA.
This study used a descriptive cross-sectional design to identify the association between falls and clinical evaluation factors. We recruited 407 participants with KOA, scheduled to undergo unilateral TKA at a single hospital in Japan between May 2019 and September 2021. The inclusion criteria were: 1. diagnosis of KOA, 2. ability to ambulate independently or with a T-cane at the time of pre-operative evaluation, 3. individuals who are scheduled for primary TKA, and 4. to obtain informed consent to participate in the study. The exclusion criteria were: 1. diagnosis of rheumatoid arthritis, idiopathic osteonecrosis, or foot and ankle disorders, 2. individuals between 60 and 84 years of age, 3. individuals with bilateral toe flexion problems, neurologic diseases, or other musculoskeletal diseases that significantly impair basic movements, such as walking, and 4. those with severe depression or dementia which would hinder evaluation.
This study complied with the Declaration of Helsinki and was approved by the Research Ethics Committee of our hospital (2019-04-21-007). Details of the study protocol and aim were explained to all participants, both verbally and in writing. All study participants then signed a written consent prior to participating in the study.
We assessed the participants one day prior to TKA. Descriptive data, fall-related assessments (presence of falls and fear of falling), radiographs, and physical function data were collected from electronic medical records. Physical function measurements were performed by 14 randomly assigned physical therapists for all participants to reduce bias as much as possible. The surgical side was considered as the affected side, whereas the non-surgical side was considered as the unaffected side.
Descriptive data, such as gender, age, height, weight, and body mass index were collected by nurses during the evaluation. The patients self-reported whether and how many times they experienced falls in the past year. 3.4 The severity of KOA was determined using the Kellgren–Lawrence (K-L) grading system [23]. Four orthopedic surgeons evaluated all medical records and determined the K-L grade severity.
A toe grip dynamometer (T.K.K.3362; Takei Scientific Instruments, Niigata, Japan) was used to measure TGS (Fig. 1 in a sitting position with 90° hip and knee joint flexion and the ankle in a neutral position. Under vertical loading on the foot, the plantar aponeurosis was extended with the foot truss structure [24]. The participants were instructed to place their test foot within the heel stopper and to grasp the dynamometer grip bar with their toes. They first performed a few test contractions with maximum effort to familiarize themselves with the measurement process and then performed as many voluntary isometric contractions as possible. Maximum TGS was measured twice, and the mean value (kg) was calculated. Participants performed maximum-effort contractions after the "warm-up" repetitions. An almost perfect inter- and intra-rater reliability of this measurement protocol using the toe grip dynamometer was observed in people aged 60–79 years [25]. Isometric knee extension strength (IKES) was measured using a hand-held dynamometer (µ-tas F1, ANIMA, Tokyo, Japan) with participants in a seated position and the knee in 90° flexion [26]. The reliability and validity of this measurement method have been previously demonstrated in patients with KOA [26, 27]. The participants were instructed to increase the intensity of knee extension against the dynamometer gradually and for approximately 2 seconds to avoid explosive contraction, and to maintain their maximal force output for approximately 3 seconds. The average of the two IKES measurements was divided by body weight (kg). Pain levels at rest and during walking were determined using a visual analog scale ranging from 0 (no pain) to 100mm (worst pain) [28]. This measurement method has been reported to be reliable and valid for the assessment of individuals with KOA [29]. The TUG was used as a behavioral measure of knee function using standard test methods [30]. The reliability and validity of this measurement method have been demonstrated [31, 32]. It was measured as the time (s) to stand up from a chair at a height of 45 cm, walk with a pole for 3 m in front of the examiner, and turn around and sit back on the chair; participants could walk with or without a T-cane. The TUG measurements were recorded twice and a mean value between the two measurements was calculated and used for statistical analysis. Fear of falling was assessed using the Japanese version of the modified Fall Efficacy Scale (mFES), developed by Hill et al [33]. The mFES is a 10-grade scale comprising 14 items (score range: 0–140 points), and a modified version of the Falls Efficacy Scale developed by Tinetti et al [34]. The mFES is used to determine the level of confidence in performing specific movements and actions without falling, with higher scores indicating higher levels of self-efficacy in fall prevention and less fear of falls. The reliability and validity of this measurement method have been demonstrated [33].
Descriptive data were presented as the number of cases, mean with standard deviations (SD), and percentages. For all analyses, the significance level was set at 5%. All statistical analyses were performed using SPSS Statistics for Windows, version 26.0 (IBM Corp, Tokyo, Japan). The sample size was calculated by G-power Post-hoc. The sample sizes for both the fall and non-fall groups were calculated based on effect size = 0.5 and α = 0.05.
Before comparing the difference between the two groups, a Kolmogorov-Smirnov test was performed as a homogeneity test. The results confirmed that the p-values for all factors were less than 0.05 for both groups. Therefore, Mann–Whitney- and chi-squared tests were used to test the difference between the two groups. The factors subjected to Mann–Whitney U test were age, BMI, TGS on both sides, IKES on both sides, pain at rest on both sides, pain while walking on both sides, TUG, and mFES. Factors subjected to the chi-squared test were sex and K-L grade.
Multiple logistic regression was performed to examine the association of each factor to the dependent variable (i.e., fall or no fall history). The independent variables were height, weight, TGS on both sides, and mFES. In addition, the variance inflation factor (VIF) was calculated to account for the degree of multicollinearity among the related factors.
The association between age and TGS in Japanese women with KOA has been previously reported [35]. Multiple logistic regression analysis of this study was performed under the same conditions described above, after propensity score matching to equalize age.
We excluded 77 participants, while 330 met the inclusion criteria (Fig. 2). The 330 participants were divided into two groups: fall and non-fall. The participants’ descriptive characteristics and results of the unpaired t-test and chi-squared test are summarized in Table 1. Of the 330 participants enrolled in the study, 74 (22.4%) were in the fall group and 256 (77.6%) were in the non-fall group. Of the participants with a fall history, 42 reported 1 fall, 14 reported 2 falls, 10 reported 3 falls, 4 reported 4 falls, and 4 reported 5 falls. A history of 6 or more falls was never recorded. The mean number of falls per participant was 1.85. When the sample size was calculated with 74 patients in the fall group and 256 in the non-fall group, the power was 0.97.
Factors that differed significantly between the fall and non-fall groups were height (p = .014), TGS on the affected side (p = .003), TGS on the unaffected side (p = .007), and mFES (p = .001).
The results of multiple logistic regression analysis are presented in Table 2. The model chi-squared test revealed significant results, indicating association with TGS on the affected side (β=-.081, p = .024, Odds ratio [OR] = .922), and mFES (β=-.013, p = .002, OR = .987). The model χ2 was significant at p < .01. The result of the Hosmer–Lemeshow test was not significant at p = .56, and the analysis result was not inconsistent. The rate of accurate discrimination was 77.6%. The predicted values did not exceed ± 3 SD relative to the measured values. VIF for mFES and TGS on the affected side were 1.023 and 1.023, respectively, and no multicollinearity was observed. VIF calculated from the related factors were height: 1.165, weight: 1.059, and TGS on the unaffected side: 2.320.
The results of multiple logistic regression analysis after propensity score matching are presented in Table 3. The results were significant in the model chi-squared test, which was associated with mFES (β=-.010, p = .050, OR = .990). The model χ2 was significant at p < .01. The result of the Hosmer–Lemeshow test was not significant (p = .18), and the analysis result was not inconsistent. The rate of accurate discrimination was 56.8%. Predicted values did not exceed ± 3 SD relative to the measured values.
We analyzed a number of factors that potentially contributed to a history of falls in patients with KOA, including TGS, and clarified which of these had an impact. Only a few studies reported the incidence of falls in individuals with KOA in Japan [36]. To the best of our knowledge, this is the first study to investigate the relationship between history of falls and TGS in older adults with KOA. Of the 5,062 frail older adults in Japan, approximately 30% experienced a fall at least once in a year [37]. Compared to older adults with KOA in Australia (48% fell within 12 months prior to TKA) [4] and in the United Kingdom (24% fell within 3 months prior to TKA) [5], the fall rate 12 months prior to TKA in our study was lower (22.4%). However, according to the publication of vital statistics in Japan, accidental deaths from falls among older adults are on the rise [38]. In addition, a previous study has reported that fall history before performing TKA increased the risk of post-operative falls.4.6 This insight might lead to a better understanding of injury prevention from post-operative falls.
The t-test identified significant differences in TGS and mFES (on the affected and unaffected sides) between the fall and non-fall groups. Adults with KOA with more frequent falls may have a more pronounced fear of falling than those with fewer falls. Tinetti et al. [39] defined fear of falling as "anxiety about falling that causes one to avoid activities of daily living, even though one is capable of performing them”. In their study, fear of falling depended on the history of falls, ranging from 12–65% among community-dwelling older adults without a fall history, and from 29–92% among those with [40]. Therefore, fall history is associated with fear of falling [41]. Our results support those of previous studies [41]. In addition, participants with KOA and a history of falls experienced knee pain, knee instability, and muscle weakness in the lower extremities, which could have also contributed to a greater fear of falling. Post-fall syndrome, in which a loss of self-confidence after a fall leads to a decreased activity level, in turn amplifies the fear of falling.
Multiple logistic regression analysis revealed that TGS on the affected side and mFES were factors related to the presence or absence of falls. A previous study has reported that TGS declines with age [42], resulting in diminished walking ability and static balance, which may be the risk factors for falls. Tsuyuguchi et al. recruited middle-aged adults and provided the age averages of all individuals (62.02) as well as a group subdivision into high and low risk of falls. They found TGS to be an independent risk factor for fall occurrence [43]. However, there is no report on whether reduced TGS in individuals with KOA is associated with the fall itself. Based on the results of our study, we believe that TGS contributes to the challenges faced by older adults with KOA and a fall history. Although a detailed causal relationship is unknown, multiple regression analysis has identified TGS as an independent factor associated with KOA [35]. Conversely, abnormal loading of the knee joint can be caused by changes in the kinematic relationship between the foot and knee [44, 45]. Compared to healthy older adults, those with KOA have lower TGS, and the measured pressure decreases during walking [46, 47]. It is possible that the progression of KOA leads to decreased TGS; conversely, decreased TGS may contribute to KOA progression. However, this causal relationship is unclear; therefore, further studies are required to investigate it. Regardless of the causal pathway, there is interdependence between TGS and KOA which increases the risk of falls. In the future, studies should approach the causal relationship between KOA and TGS from the perspectives of kinesiology and biomechanics, in addition to seeking strategies to prevent falls. Another interesting finding is that an increase in TGS reduced knee pain and knee extensor strength [7], which have been reported to be associated with falls in individuals with KOA. The importance of assessing TGS in addition to existing assessments in individuals with KOA has been emphasized.
In this study, propensity score matching was performed to equalize age before performing logistic regression analysis. As a result, TGS was no longer a relevant factor, and only mFES was extracted as a relevant factor. As expected, age equalization led to the exclusion of TGS from the logistic regression analysis. Age itself was not significantly different between men and women with and without a history of falls. On the other hand, mFES was extracted as an associated factor even after age equalization. Even after adjusting for age, participants with a fall history had fear of falling, while participants without did not.
This study had three major limitations. First, TGS was the only assessment performed on the foot. The degree of flatfoot and range of motion of the foot, which are common problems in individuals with KOA, were not measured. The abnormal foot posture of KOA has excessive first medial tibiofemoral contact force during walking [48]. As these factors are also associated with gait, which is involved in half of all fall scenarios [49], they are likely to contribute to falls. Second, previous studies have reported decreased physical activity [50] as well as hip [51] and ankle [52] weakness as risk factors for falls. Because we did not evaluate these factors and did not include these results in our regression analysis, we could not determine the relative contribution of TGS to physical activity and muscle strength for the above factors in older adults with KOA. Finally, we conducted a cross-sectional observational study; therefore, subsequent cohort and intervention studies should be conducted to better clarify the causal relationships between foot function and falls in older adults with KOA.
In this study, we investigated the factors leading to falls in older adults with KOA in Japan. Our results indicated that falls in older adults with KOA were related to TGS as well as increased age. Physical therapy interventions to enhance TGS could be one solution to help prevent falls in individuals with KOA. In the future, cohort and interventional studies evaluating the relationship between TGS and falls should be performed.
Confidence interval
Isometric knee extension strength
Knee osteoarthritis
modified Fall Efficacy Scale
Odds ratio
Standard deviation
Toe grip strength
Total knee arthroplasty
Timed Up and Go
Variance inflation factor
Ethics approval
The study protocol was approved by the institutional review boards of Kashiba Asahigaoka Hospital (2019-04-21-007). This study complied with the Declaration of Helsinki. Details of the study protocol and the aim of the study were explained to all participants, and written informed consent was obtained from all participants before including them in the study.
Consent for publication
Not applicable.
Availability of data and materials
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.
Competing interests
The authors declare that they have no competing interests.
Funding Statement
Our study was not funded by grants, research programs, or institutional support.
Author Contributions
Conceptualization: Yuya Mawarikado, Yusuke Inagaki, Akira Kido, Yasuhito Tanaka
Data curation: Yuya Mawarikado, Yusuke Inagaki
Formal analysis: Yuya Mawarikado, Yusuke Inagaki
Investigation: Yuya Mawarikado, Yusuke Inagaki, Takanari Kubo
Methodology: Yuya Mawarikado, Yusuke Inagaki, Takanari Kubo
Project administration: Yuya Mawarikado
Resources: Yuya Mawarikado, Tadashi Fujii
Software: Yuya Mawarikado, Yusuke Inagaki
Visualization: Yuya Mawarikado
Writing – original draft: Yuya Mawarikado, Yusuke Inagaki, Akira Kido, Yasuhito Tanaka
Acknowledgements
We would like to thank Editage (www.editage.jp) for English language editing.
Tables 1 to 3 are available in the Supplementary Files section.