To examine associations of foot symptoms with self-reported and performance-based measures of physical function in a large, biracial, community-based sample of individuals ages ≥45 years.
To examine associations of foot symptoms with self-reported and performance-based measures of physical function in a large, biracial, community-based sample of individuals ages ≥45 years.
Data from 2,589 Johnston County participants (evaluated in 1999–2004) were used in cross-sectional analyses. The presence of foot symptoms was defined as pain, aching, or stiffness of at least one foot on most days. Physical function was assessed by the total Stanford Health Assessment Questionnaire (HAQ) score (0, >0 but <1, and ≥1), timed 5 repeated chair stands (completion time <12 seconds, ≥12 seconds, and unable), and 8-foot walk time (<3.35 seconds and ≥3.35 seconds). Separate multivariable logistic regression models examined associations between foot symptoms and physical function measures, controlling for age, race, sex, body mass index, radiographic knee osteoarthritis, radiographic hip osteoarthritis, knee symptoms, hip symptoms, and depressive symptoms. Interaction terms between each of the 3 physical function measures and each demographic and clinical characteristic were examined.
The prevalence of foot symptoms was 37%. Participants with foot symptoms were more likely than those without symptoms to have higher HAQ scores (adjusted odds ratio [OR] 1.79, 95% confidence interval [95% CI] 1.50–2.12). Among obese participants, those with foot symptoms had longer chair stand (adjusted OR 1.38, 95% CI 1.04–1.87) and 8-foot walk times (adjusted OR 1.61, 95% CI 1.21–2.15) than those without symptoms.
Foot symptoms were independently and significantly associated with 2 of 3 measures of poorer physical function. Interventions for foot symptoms may be important for helping patients prevent or deal with an existing decline in physical function.
Foot pain is common among older adults, affecting approximately 20–30% of community-dwelling adults ages 65 years and older (1–3). An association between foot pain and decreased functional abilities has been demonstrated in several studies (2–8). Most prior studies of foot symptoms (i.e., pain, aching, and stiffness) and functional abil-ities have focused on adults ages 62 years and older (2, 3, 5, 6, 8), but this relationship has been explored less in middle-aged adults (9, 10) and important public health subsets, such as obese adults. Furthermore, few studies have examined these associations controlling for the effects of osteoarthritis (OA). Identifying adults at higher risk of functional limitation related to foot symptoms ear-lier in the life course may aid in preventing or slowing functional loss. Additionally, prior investigations of the association between foot symptoms and disability have focused primarily on self-report or physical assessments of lower extremity function (3, 5, 6, 8). However, both lower and upper extremity task performance may be affected by lower extremity symptoms, as illustrated by a prior study of knee pain and lower and upper extremity function (11).
The purpose of the present study was to assess the association between foot symptoms and self-reported and performance-based measures of physical function in a large, biracial, community-based sample of individuals that includes younger individuals (age 45 years and older) than the populations usually studied. These associations were examined to determine whether they were independent of knee and hip radiographic OA and knee and hip symptoms, and whether these associations varied by age, race/ethnicity, sex, body mass index (BMI), and depressive symptoms.
This cross-sectional sample was composed of individuals enrolled in the Johnston County Osteoarthritis Project, an ongoing, community-based study of the occurrence of knee and hip OA in African American and white residents in a rural county in North Carolina. Details of this study have been reported previously (12). Briefly, this study involved civilian, noninstitutionalized adults ages 45 years and older who resided in 6 townships in Johnston County. Participants were recruited by probability sampling using 1990 census data, with oversampling of African Americans. The sampling design consisted of selection of primary sampling units by stratified simple random sampling of streets, and households were counted on each street. Age, sex, ethnic group, and marital status of each household were collected, and age and ethnicity information was used to determine the eligibility of each household. A total of 14,297 households were identified, and 4,866 were eligible. From May 1991 to December 1998, first home baseline interviews were completed by 3,690 participants (76% of the eligible 4,866), and 83% (3,068 participants) of those who completed the interview were examined at the local clinic.
The present analysis used data collected from individuals who had participated in the clinic visit arms of either the first followup study (n = 1,699) in 1999–2004, or the new enrollment study (n = 1,007) in 2003–2004 (total study sample = 2,706). The new enrollment aimed to enrich the sample for African Americans and younger individuals; therefore, these participants were younger (mean age 59.6 versus 66.0 years) and more likely to be African American (39.9% versus 27.9%), as compared to participants in the first followup study.
Participants completed an interviewer-administered questionnaire in which they answered “Yes” or “No” separately for their right and left feet: “On most days, do you have pain, aching, or stiffness in your right/left foot?” Participants were considered to have foot symptoms if they answered affirmatively to the foot symptoms question for at least one foot.
Participants were asked separately for left and right knees and left and right hips: “On most days do you have pain, aching, or stiffness in your [left/right] [knee/hip]?” Participants were considered to have knee symptoms if they answered affirmatively to the knee symptoms question and were considered to have hip symptoms if they answered affirmatively to the hip symptoms question.
All of the participants completed bilateral anteroposterior radiography of the knee in weight bearing. Women ages >50 years and all of the men completed supine anteroposterior pelvic radiography. Radiographs were rated by a single musculoskeletal radiologist (JBR) using the Kellgren/Lawrence radiographic atlas for overall knee and hip radiographic grades (13). As previously described, interrater reliability (comparison of radiograph readings between JBR and another radiologist) and intrarater reliability (comparison of radiograph readings completed by JBR at two separate times) for the radiologist were high (weighted κ for interrater reliability = 0.9, κ for intrarater reliability = 0.9) (12). Using the Kellgren/Lawrence scale, radiographs were classified as showing radiographic OA if they had a grade of 2–4 (showing an osteophyte or having joint space narrowing) (14). Associations between foot pain and radiographic OA in areas other than the foot have been shown in previous studies (5, 7).
Self-reported functional status was assessed with the Stanford Health Assessment Questionnaire (HAQ) Disability Index (15). Twenty individual functions covering 8 domains (dressing, arising, eating, walking, reaching, gripping, chores, and hygiene) were scored on a scale of 0–3 (where 0 = with no difficulty, 1 = with little difficulty, 2 = only with much difficulty, and 3 = unable to do). A score of 2 was assigned for an individual function if the participant needed an assistive device to perform that activity. The highest score of the individual functions in each domain determined the score for that domain. HAQ scores were calculated by averaging the scores of the 8 domains. HAQ scores were available for 2,581 participants with foot symptoms data (Figure 1).
Performance-based functional status was assessed using timed 5 chair stands and a timed 8-foot walk (16). The chair stand test is considered an assessment of lower extremity strength, and the 8-foot walk assesses gait velocity, a component of walking functional ability (17). Although most studies have assessed the validity of the chair stand test among elderly adults, one study of adults ages 55–70 years reported good validity of a chair stand test for assessing lower extremity strength when compared to a 1 repetition maximum leg press (Pearson's r = 0.68, P < 0.05) (18). Gait velocity has been examined across all ages of adults, and walking speed does decline with age (19). Participants were asked to rise from a chair of standard height consecutively 5 times as quickly and comfortably as possible without using their arms while a study evaluator measured performance time with a stop watch. If a participant completed fewer than 5 stands, the performance was recorded as “unable.” If a participant completed 5 stands, the time of completion to the nearest tenth of a second was recorded. For the timed 8-foot walk, participants completed two timed trials of walking 8 feet at their usual walking pace and using walking aids if needed. Times were measured with a stopwatch and recorded to the nearest tenth of a second for each trial. For analyses, the timed 8-foot walk was calculated as the average of the two trials.
The following participant characteristics were examined as potential covariates in our analyses because they may be associated with foot symptoms and physical function: sex, self-reported race (African American or white), age (continuous variable in years), BMI at baseline (continuous variable calculated as weight in kilograms/height in meters squared), and depressive symptoms (reported using the Center for Epidemiologic Studies Depression Scale [CES-D] ). Menz and Morris (6) reported that female sex and higher BMI were each strongly associated with self-reported disabling foot pain. Height without shoes was measured in centimeters and weight was measured in kilograms using a balance beam scale.
Cochran-Mantel-Haenszel statistics for categorical variables and one-way analysis of variance testing of linear trend for continuous variables were used to compare all demographic and clinical characteristics (sex, race, age, BMI [continuous variable], knee radiographic OA, hip radiographic OA, knee symptoms, hip symptoms, and CES-D) across the 3 levels of the HAQ score. Three functional outcomes were considered: total HAQ score, timed 5 chair stands, and timed 8-foot walk. The total HAQ score was categorized into 3 levels (0, >0 but <1, and ≥1). The median value of each physical performance measure was used as a cut point to create categorical variables. The timed 5 chair stands were coded as a 3-level variable (median completion time of <12 seconds [referent], ≥12 seconds, and unable). The timed 8-foot walk was coded as a dichotomous variable, setting the cut point at the median completion time (<3.35 seconds and ≥3.35 seconds). Participants with missing exposure or outcome data were excluded from analyses (Figure 1).
Associations between foot symptoms and each physical function measure were examined in separate logistic regression models, adjusting first for age, sex, race, and BMI, then additionally for knee and hip symptoms and knee and hip radiographic OA, and finally for CES-D score. The proportional odds assumption was tested for the 3-level outcomes of the HAQ and the timed 5 chair stands. A proportional odds model assumes that the relationship between the independent variable of interest and the 3 levels of outcome is similar across successively more severe levels of the outcome. A single odds ratio (OR) was calculated to describe the comparison between 1) the best physical function (referent) category (e.g., HAQ = 0) and the 2 other categories combined and 2) the best physical function category plus the middle category (e.g., 0 < HAQ < 1) compared with the worst category (e.g., HAQ ≥1). A generalized logit model was used if the proportional odds assumption was violated. Therefore, a proportional odds model was used to analyze associations for the HAQ score outcome, and a generalized logit model was used for analyses of the timed 5 chair stands. A generalized logit model generates an OR for each comparison between the referent (<12 seconds) and the middle physical function category (≥12 seconds) and between the referent and the worst physical function category (unable). Interactions between the foot symptom variable and race, sex, BMI (categorized as >30 kg/m2 [obese] and ≤30 kg/m2 [non-obese] for analyses of interactions), hip symptoms, and knee symptoms were tested jointly in each model. In models with significant interactions (P values less than 0.05), appropriate subgroups were examined, and ORs and 95% confidence intervals (95% CIs) were calculated separately for each subgroup. The Hochberg-Benjamini method (21) was used to adjust for multiple comparisons for each outcome. All statistical computations were performed using SAS software, version 9.1 (SAS Institute).
Of the initial 2,706 participants clinically evaluated, all had data available for analyses, but 5 (0.2%) with evidence of an inflammatory arthropathy on knee or hip radiographs were excluded, leaving 2,701 for analysis (Figure 1). Foot symptoms data were available for 2,589 (95.9%) of 2,701 participants (Figure 1), and chair stand and 8-foot walk data were available for 2,572 (99.3%) of 2,589 participants with foot symptoms data (Figure 1). Foot symptoms were reported by 37.0% (n = 957) of participants. Knee symptoms were reported by 50.1% (n = 1,296) and hip symptoms were reported by 38.3% (n = 991). Radiographic knee OA was present in 34.8% (n = 901) of participants, and 30.6% (n = 793) had radiographic hip OA. The characteristics of most variables of the study population are shown in Table 1. Obesity, defined as a BMI >30 kg/m2, was present in 46.6% (n = 1,206) of participants. Approximately 18.8% (n = 484) were unable to perform the timed 5 chair stands, and the median completion times for 5 chair stands and the 8-foot walk were 12 seconds and 3.35 seconds, respectively. Compared to participants without foot symptoms, those with foot symptoms were more likely to have a higher BMI, be women, have radiographic knee OA, have knee or hip symptoms, report higher HAQ scores, have slower chair stand and 8-foot walk completion times, and report depressive symptoms (Table 1).
|Total, no. (%)||2,589 (100)||957 (37.0)||1,632 (63.0)|
|Age, mean ± SD years||63.2 ± 10.4||63.4 ± 10.4||63.3 ± 10.5||0.8264|
|BMI, mean ± SD kg/m2||30.6 ± 6.7||31.9 ± 7.0||29.8 ± 6.4||< 0.0001|
|African American||843/2,589 (32.6)||297/957 (31.0)||546/1,632 (33.4)||0.2083|
|Women||1,686/2,589 (65.1)||670/957 (70.0)||1,016/1,632 (62.3)||< 0.0001|
|Radiographic hip OA||793/2,589 (30.6)||311/957 (32.5)||482/1,632 (29.5)||0.1190|
|Radiographic knee OA||901/2,589 (34.8)||372/957 (38.9)||529/1,632 (32.4)||0.0008|
|Hip symptoms||991/2,588 (38.3)||536/957 (56.0)||455/1,631 (27.9)||< 0.0001|
|Knee symptoms||1,296/2,588 (50.1)||669/957 (69.9)||627/1,631 (38.4)||< 0.0001|
|HAQ score||< 0.0001|
|≥1||571/2,581 (22.1)||344/952 (36.1)||227/1,629 (13.9)|
|>0 but <1||935/2,581 (36.2)||372/952 (39.1)||563/1,629 (34.6)|
|0||1,075/2,581 (41.7)||236/952 (24.8)||839/1,629 (51.5)|
|Chair stand||< 0.0001|
|Unable||484/2,572 (18.8)||271/949 (28.5)||213/1,623 (13.1)|
|≥12 seconds||875/2,572 (34.0)||329/949 (34.7)||546/1,623 (33.6)|
|<12 seconds||1,214/2,572 (47.2)||349/949 (36.8)||865/1,623 (53.3)|
|8-foot walk||< 0.0001|
|≥3.35 seconds||1,285/2,572 (50.0)||563/949 (59.3)||722/1,623 (44.5)|
|<3.35 seconds||1,287/2,572 (50.0)||386/949 (40.7)||901/1,623 (55.5)|
|CES-D score, mean ± SD||6.6 ± 8.5||8.5 ± 9.5||5.5 ± 7.7||< 0.0001|
Those with higher HAQ scores were more likely to be older adults, women, or African American, and to have a higher mean BMI, radiographic hip or knee OA, knee or hip symptoms, or depressive symptoms (Table 2). These same groups required longer times to complete the timed 5 chair stands and timed 8-foot walk (data not shown).
|Covariate||HAQ = 0 (n = 1,075 [41.7%])||0 < HAQ < 1 (n = 935 [36.2%])||HAQ ≥1 (n = 571 [22.1%])||P|
|Age, mean ± SD years||62.7 ± 9.9||63.4 ± 10.4||64.5 ± 11.3||< 0.0001|
|Age groups, years||< 0.0001|
|44 to <55||280 (26.1)||237 (25.3)||135 (23.6)|
|55 to <65||371 (34.5)||303 (32.4)||186 (32.6)|
|65 to <75||283 (26.3)||253 (27.1)||126 (22.1)|
|≥75||141 (13.1)||142 (15.2)||124 (21.7)|
|African American||307 (28.6)||306 (32.7)||228 (39.9)||< 0.0001|
|Women||612 (56.9)||622 (66.5)||445 (77.9)||< 0.0001|
|BMI, mean ± SD kg/m2||29.2 ± 5.4||30.7 ± 6.7||33.1 ± 8.0||< 0.0001|
|Radiographic hip OA||307 (28.6)||281 (30.0)||202 (35.4)||0.0037|
|Radiographic knee OA||279 (26.0)||346 (37.0)||274 (48.0)||< 0.0001|
|Hip symptoms||216 (20.1)||387 (41.4)||382 (66.9)||< 0.0001|
|Knee symptoms||308 (28.7)||522 (55.8)||460 (80.6)||< 0.0001|
|CES-D score, median ± SD||3.8 ± 5.9||6.6 ± 7.7||12.1 ± 11.1||< 0.0001|
Participants with foot symptoms were more likely than those without symptoms to have higher total HAQ scores after adjusting for the study covariates (adjusted OR 1.79, 95% CI 1.50–2.12) (Table 3). Statistically significant interactions were observed between foot symptoms and BMI for the logistic regression models of chair stand and 8-foot walk outcomes. Among obese (BMI >30 kg/m2) participants, those with foot symptoms required more time to complete the chair stands (unable versus <12 seconds: adjusted OR 2.12, 95% CI 1.46–3.07; ≥12 seconds versus <12 seconds: adjusted OR 1.38, 95% CI 1.04–1.87) and the 8-foot walk (adjusted OR 1.61, 95% CI 1.21–2.15) than those without symptoms, a difference not seen for non-obese participants (chair stands: adjusted OR 1.21, 95% CI 0.81–1.79 for unable versus <12 seconds and adjusted OR 0.88, 95% CI 0.66–1.18 for ≥12 seconds versus <12 seconds; 8-foot walk: adjusted OR 1.13, 95% CI 0.86–1.50) (Table 4).
|Foot symptoms, adjusted OR (95% CI)||P|
|Crude||3.34 (2.86–3.90)||< 0.001|
|Adjusted 1*||3.05 (2.61–3.57)||< 0.001|
|Adjusted 2†||1.95 (1.65–2.31)||< 0.001|
|Adjusted 3‡||1.79 (1.50–2.12)||< 0.001|
|Outcome||BMI ≤30 kg/m2 (53.4%), adjusted OR (95% CI)||BMI >30 kg/m2 (46.6%), adjusted OR (95% CI)||P for interaction|
|Timed 5 chair stands, seconds|
|Unable vs. <12 (referent)||1.21 (0.81–1.79)||2.12 (1.46–3.07)||0.0110|
|≥12 vs. <12†||0.88 (0.66–1.18)||1.38 (1.04–1.87)||0.0290|
|Timed 8-foot walk, seconds|
|≥3.35 vs. <3.35 (referent)||1.13 (0.86–1.50)||1.61 (1.21–2.15)||0.0314|
Thirty-seven percent of participants in the present study reported foot symptoms, slightly higher than the proportion reported in studies of adults ages 65 years and older, signifying that foot symptoms are also a common condition in non-elderly adults. Participants with foot symptoms were more likely than those without foot symptoms to report greater difficulty with self-reported and performance-based physical function (as indicated by the 8-foot walk), even when controlling for important covariates. These results among participants ages 45 years and older are consistent with prior studies that have examined adults ages 62 years and older, demonstrating a persistent relationship association even at younger ages (2, 3, 5, 6, 8).
The measure of self-reported physical function, the HAQ, includes questions about both upper and lower extremity function. A direct relationship between foot symptoms and lower extremity function can be easily appreciated. However, foot symptoms may affect one's overall confidence in his or her ability to complete physical tasks (functional self-efficacy). Therefore, inclusion of a measure such as the HAQ was important to attempt to reflect the functional self-efficacy associated with foot symptoms. Self-efficacy was not assessed specifically in the present study, but the results were independent of depressive symptoms, which have been shown to contribute to self-efficacy for general physical tasks (22).
Foot symptoms may affect physical function by contributing to altered gait patterns, ultimately reducing gait speed. This may explain the slower completion times noted among participants with foot symptoms during the 8-foot walk. In general, no association was observed between foot symptoms and timed chair stands comparing ≥12 to <12 seconds. However, the presence of obesity positively modified this association; obese participants (BMI >30 kg/m2) with foot symptoms required more time to complete the chair stand task than those without symptoms, but this association was not observed among non-obese participants with and without foot symptoms. Therefore, weight reduction may assist in diminishing the impact of foot symptoms on some functional activities.
Strengths of this study include that it is community-based, consists of African American and white men and women, and includes radiographic OA data and knee and hip symptoms data. Additionally, to our knowledge, this is the largest study to date that includes assessment of foot symptoms and self-reported and physical function. One limitation of the present study is that conditions contributing to foot symptoms in this population are not known. Foot radiographs were not obtained, but it is unlikely that participants had arthritic and rheumatic conditions other than OA, because participants with inflammatory conditions on knee or hip radiographs were excluded. Other conditions that may have contributed to foot symptoms, such as calluses, corns, hallux deformities, toe deformities, pes planus, heel spurs, plantar fasciitis, and neuropathy, were not included in these analyses, but will be the topic of future research. A comprehensive physical assessment of the foot may help identify musculoskeletal origins of foot symptoms in this population, but previous studies have demonstrated that some foot disorders, such as hallux valgus and toe deformities, and especially mild disease, may not be associated with foot symptoms or functional limitations (5, 23). Another possible limitation is that results for women are only generalizable to women over the age of 50 years because hip radiographs were obtained only for women who were at least 50 years of age, due to radiation concerns with reproductive health. However, in contrast to prior comparable studies (2, 5), the present study included middle-aged adults (at least 45 years of age). Additionally, whether the foot symptoms were acute or chronic in this population is not known because the duration of foot symptoms was not collected during study interviews. Furthermore, a measure of self-reported physical function that was specific to the foot was not collected at baseline. Finally, due to the cross-sectional study design, a temporal relationship between foot symptoms and physical function cannot be inferred. A longitudinal analysis would assist in determining the possible contribution of foot symptoms to functional decline.
In summary, foot symptoms play an important role in poor physical function, independent of knee and hip symptoms and OA. Interventions for foot symptoms, including weight loss, foot orthotics, foot care strategies, and patient education on proper footwear, may be important for helping patients prevent or manage an existing decline in perceived and performance-based functional abilities. Screening strategies may be helpful in identifying patients with foot symptoms who are at risk of the development or progression of functional decline.
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. Golightly 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. Golightly, Hannan, Shi, Helmick, Renner, Jordan.
Acquisition of data. Renner, Jordan.
Analysis and interpretation of data. Golightly, Hannan, Shi, Helmick, Renner, Jordan.