Volume 46 Number 9, 2009
   Pages 1091 — 1098

Relationship between depression and functional measures in overweight and obese persons with osteoarthritis of the knee

Daniel Possley, DO;1-2 Elly Budiman-Mak, MD;3-4 Susan O'Connell, MHA, RN;1 Christine Jelinek, MS;1 Eileen G . Collins, PhD, RN1,3,5*

Abstract — Depression is common in overweight and obese individuals with chronic illness. The purpose of this study was to determine the relationship between depression and functional status. Baseline data were used from 105 overweight/obese participants who enrolled in a clinical trial for overweight and obese adults with osteoarthritis of the knee. Forty-two percent of the sample was classified as depressed according to the Center for Epidemiologic Studies Depression Scale. A moderate relationship was seen between perceived physical function and physical performance in patients who were not depressed that did not exist in patients reporting depressive symptoms. In a stepwise regression analysis, poorer function (as measured by the Western Ontario and MacMaster Universities (Osteoarthritis Index) function subscale) and younger age accounted for 29 percent of the variance in depressive symptoms.

Key words: 6-minute walk, depression, obese, osteoarthritis, osteoarthritis knee, overweight, physical activity, physical function, quality of life, walking impairment.

Osteoarthritis (OA) is the most common form of arthritis, affecting an estimated 21 million people in the United States in 1990 [1]. Depression is also highly prevalent in the United States, with 19 million people over the age of 18 years being affected by a depressive disorder in any given year [2]. In a recent longitudinal study of OA sufferers, 25.2 percent of 369 subjects met the criteria for depression [3-5].

The purpose of this study was to examine the relationship between depression and functional status of overweight and obese patients with OA of the knee. Our objectives were to (1) examine the prevalence of depressive symptoms in our sample; (2) examine the relationship between depressive symptoms, perceived physical function, and physical performance in overweight and obese adults with OA of the knee; and (3) determine what factors were associated with depressive symptoms. We hypothesized that (1) patients who reported more functional limitations would also report more depression, and (2) lower reported function will account for significant variance in depressive symptoms for this sample.

After the study was approved by the Human Studies Subcommittee, we recruited patients with documented OA of the knee from the General Medicine and Arthritis clinics at a midwestern Department of Veterans Affairs (VA) Hospital and the surrounding community. Patients were identified with International Classification of Diseases-9th Revision codes 715.06 or 715.26 as a primary or secondary diagnosis. Identified patients were sent a letter of invitation. To be contacted by the staff, the patient needed to respond positively to the letter of invitation by returning a stamped, self-addressed information sheet. We then invited patients for an initial visit in the laboratory and obtained written informed consent.

Patients 50 years or older with documented OA of the knee were invited to participate. We documented OA of the knee as radiographic findings of tibiofemoral joint changes of Grades 2 to 3 using the Kellgren-Lawrence criteria. Additionally, we included patients if their knee pain was symptomatically defined as mild to severe pain for most days of the week for at least 6 months before entry and if they were functional Class I to III of the American Rheumatology Association functional capacity criteria, had a body mass index (BMI) between 27 and 40 kg/m2, and were able to attend scheduled exercise and nutrition counseling sessions.

Patients were excluded if they had undergone arthroscopic surgery of the knee within the previous 3 months; had severe back, hip, or ankle pain that prevented them from walking; or had joint replacements in both knees.

The Center for Epidemiologic Studies Depression Scale (CES-D) is a 20-item measure designed to evaluate symptoms of depression in the community population [6]. A score of £16 indicates that a person is not depressed, and a score of >16 indicates that a person is depressed. The CES-D has been shown to be reliable in general and patient populations, with internal consistency reported at 0.85 to 0.90 [6], and valid in patients with arthritis [7]. The CES-D has been successfully used in depression evaluation [6,8] and in normal and psychiatric populations and correlates with the clinical diagnosis of depression [6].

The Western Ontario and MacMaster Universities (Osteoarthritis Index) (WOMAC) Likert 3.1 questionnaire is designed to measure functional loss and pain in persons with arthritis [9-10]. In the difficulty domain, the focus is measuring tasks related to activities of daily living. The WOMAC is a disease-specific measure of factors affecting quality of life in persons with OA. It contains 17 functional activity items, 5 pain-related items, and 2 stiffness items. The WOMAC is a valid and reliable instrument that has been used widely in a variety of clinical trials involving patients with OA of the knee [9,11-14].

The Walking Impairment Questionnaire (WIQ) is a self-administered, 14-item survey with three subscales: walking distance (7 items), walking speed (4 items), and stair-climbing ability (3 items). The WIQ assesses subjects' self-reported difficulty in walking a defined distance (half or less, one, two, or three city blocks); difficulty with which and how fast they are able to walk one block (walking one block slowly, at an average speed, or quickly or running or jogging one block); and difficulty climbing one, two, and three flights of stairs. Therefore, WIQ measures an individual's mobility and degree of difficulty in performing the tasks. All responses are scored on a 0 to 4 scale (4 = no difficulty, 3 = slight difficulty, 2 = some difficulty, 1 = much difficulty, and 0 = unable to do). The WIQ is scored by multiplying the degree of difficulty by a predefined number to represent distance (feet), speed (miles per hour) or stairs climbed (steps). The derived number is divided by the total possible score. A whole number percentage is then obtained by multiplying this value by 100. The WIQ is primarily used in patients with peripheral arterial disease [15-16] but has also been validated in the cohort of patients in our study [17].

The 36-Item Short Form Health Survey (SF-36) was used to measure health-related quality of life. The SF-36, widely used, is a valid and reliable measure [18]. We used the Physical and Mental Composite scores in this analysis.

To assess stair-climbing ability, we asked subjects to complete a 3-minute stair climb and descend activity. At the signal to begin, subjects climbed four steps up, encountered a small landing at the top, and descended four steps down. When the subject reached the bottom of the four stairs, they turned 180× and climbed and descended again [19]. For 3 minutes, the subject moved continuously, completing as many stairs as possible in the allotted time. Subjects were instructed that if they needed to stop before the 3 minutes were completed to simply stop where they were.

The 6-minute walk is a test in which the subject walks for 6 minutes on level ground and the distance covered is measured [20]. The walk test was carried out in a 30.48 m level indoor straight corridor of the hospital. Subjects were instructed to cover as much ground as possible in 6 minutes. Patients were told to walk continuously if possible but that it would be permissible to slow down or stop if they felt it necessary to do so. The goal of the test was to have subjects adopt a self-selected walking speed such that at the end of 6 minutes, they would feel that they could not have walked any further. Subjects were encouraged and told when they had walked 1, 3, and 5 minutes.

We used descriptive statistics (mean, standard deviation, range, and frequency) to characterize the sample and outcome measures. Pearson r correlations examined relationships between outcome measures. Bivariate analysis compared the demographics and clinical characteristics of the cohort with/without depression. Hierarchical linear regression procedures examined predictors of depression and physical function. We used SPSS version 15.0 (Chicago, Illinois) for our analysis.

A total of 112 patients were enrolled in the study. Of these, 7 had missing data on the questionnaires or did not complete the questionnaires. Therefore, we used data from 105 patients in this analysis.

The sample consisted primarily (89%) of older men (mean age ± standard deviation = 67.5 ± 8.4 yr) who reported having OA of the knee for an average of 10 years. The majority were Caucasian (89%) and African American (6%). They were 173.6 ± 8.3 cm tall and weighed 103.5 ± 19.5 kg. They were obese with a BMI of 34.5 ± 5.9 kg·m2. Sixty subjects (56%) were classified as Kellgren-Lawrence Class 2 and forty-five (44%) as Kellgren-Lawrence Class 3. On average, patients reported 2.3 ± 2.0 coexistent diseases with their OA of the knee.

Mean scores for the CES-D, WOMAC, WIQ, SF-36, 6-minute walk (distance), and 3-minute stair climb and descend activities (number of stairs) are listed in Table 1. The relationship between depression and major study variables such as age, duration of OA, number of coexisting comorbidities, WOMAC scores, and WIQ scores are presented in Table 2. Using the CES-D score, we classified patients as depressed (>16) or not depressed (£16). Sixty-one patients (58%) were classified as not depressed according to the CES-D, and forty-four (42%) as depressed. Those who were classified as depressed had more coexistent diseases, more impaired walking, had OA for a longer period of time, and more pain and stiffness and poorer function than those who were not classified as depressed (Table 3). We then examined the relationship between perceived physical function and measured physical performance among the depressed and nondepressed patients. Results are presented in Table 4. Stronger relationships were seen between perceived and measured physical ability in the nondepressed patients than in the depressed patients on the WOMAC.

---

Table 1. Descriptive information on selected variables in study of overweight patients with osteoarthritis of the knee.
Measure	Mean ± SD	Possible Range
CES-D	15.4 ± 9.5	0-60
WOMAC
Pain Subscale	12.3 ± 3.5	5-25
Stiffness Subscale	5.6 ± 1.6	2-10
Function Subscale	41.7 ± 12.4	17-85
Walking Impairment Questionnaire
Distance Subscale	46.9 ± 28.7	0-100
Speed Subscale	39.5 ± 22.7	0-100
Stair Climbing Subscale	46.7 ± 26.2	0-100
36-Item Short Form Health Survey
Physical Composite Score	36.9 ± 8.3	0-100
Mental Composite Score	50.1 ± 10.9	0-100
6-Minute Walk (distance [m])	443.8 ± 102.7	0-no upper limit
3-Minute Stair Climb and Descend (stairs [No.])	176.0 ± 52.8	0-no upper limit
CES-D = Center for Epidemiologic Studies Depression Scale, SD = standard deviation, WOMAC = Western Ontario and MacMaster Universities (Osteo-arthritis Index).

---

Table 2. Relationship between depression and selected variables.
Variable	CES-D
	r-Value	p-Value
Age (yr)	-0.35	0.001
Duration of OA (yr)	0.20	0.20
Kellgren-Lawrence Classification	-0.06 (Kendall's t)	0.45
Coexisting Diseases (n)	0.34	<0.001
BMI (kg/m2)	0.13	0.10
WOMAC Function	0.54	<0.001
WOMAC Stiffness	0.26	0.004
WOMAC Pain	0.43	<0.001
WIQ Total Score	-0.46	<0.001
SF-36 Physical Composite Score	-0.28	0.002
6-Minute Walk (distance [m])	-0.17	0.04
3-Minute Stair Climb and Descend (stairs [No.])	-0.14	0.07
BMI = body mass index, CES-D = Center for Epidemiologic Studies Depression Scale, OA = osteoarthritis, SF-36 = 36-Item Short Form Health Survey, WIQ = Walking Impairment Questionnaire, WOMAC = Western Ontario and MacMaster Universities (Osteoarthritis Index).

---

Table 3. Differences between depressed versus not depressed subjects.
Variable	Not Depressed, n = 61 (Mean ± SD)	Depressed, n = 44 (Mean ± SD)
Age (yr)	69.1 ± 8.1	65.9 ± 8.6
Duration of OA (yr)	7. 7 ± 4.9	11.9 ± 11.4
BMI (kg/m2)	33.9 ± 5.3	34.8 ± 6.5
% Body Fat	39.2 ± 7.3	40.2 ± 8.2
Comorbid Diseases (n)	1.8 ± 1.6	2.9 ± 2.4
WOMAC Function	20.8 ± 9.7	30.3 ± 12.3
WOMAC Stiffness	6.5 ± 2.8	8.1 ± 3.6
WOMAC Pain	6.3 ± 2.9	8.5 ± 3.5
WIQ Total Score	66.7 ± 26.4	48.9 ± 28.0
SF-36 Physical Composite Score	37.8 ± 8.1	36.0 ± 8.3
Leg Strength (kg)	43.0 ± 15.3	40.4 ± 12.4
6-Minute Walk (distance [m])	453.3 ± 103.8	432.0 ± 103.0
3-Minute Stair Climb and Descend (stairs [No.])	182.7 ± 51.1	169.3 ± 53.6
BMI = body mass index, OA = osteoarthritis, SD = standard deviation, SF-36 = 36-Item Short Form Health Survey, WIQ = Walking Impairment Questionnaire, WOMAC = Western Ontario and MacMaster Universities (Osteoarthritis Index).

---

Table 4. Relationship (r-value) between perceived and measured physical abilities in nondepressed and depressed patients.
Measure	Nondepressed Patients			Depressed Patients
	6-Minute Walk	3-Minute Stair Climb/Descend		6-Minute Walk	3-Minute Stair Climb/Descend
WOMAC Function	-0.49*	-0.50*		-0.09	0.06
WOMAC Pain	-0.28	-0.28†		-0.06	-0.01
WOMAC Stiffness	-0.02	0.03		0.04	0.12
WIQ Distance	0.50*	0.52*		0.49*	0.33†
WIQ Speed	0.59*	0.57*		0.41‡	0.33†
WIQ Stair-Climbing	0.44*	0.51*		0.43‡	0.30
SF-36 Physical Composite Score	0.25†	0.30†		0.30	0.22
*p < 0.001. †p < 0.05. ‡p < 0.01. SF-36 = 36-Item Short Form Health Survey, WIQ = Walking Impairment Questionnaire, WOMAC = Western Ontario and MacMaster Universities (Osteoarthritis Index).

We ran hierarchical, stepwise regression to determine predictors of depression in this sample. Using only variables with a significant correlation (p = 0.10) with depression, we determined a final compressed model. The final model included age as a demographic predictor and functional status as determined by the WOMAC function subscale. Perceived function predicted 24 percent of the variance while age predicted 5 percent of the variance. Together, the model predicted 29 percent of the variance of the depression scores (Table 5). Those patients who were younger with lower perceived function demonstrated higher rates of depression.

---

Table 5. Regression analysis summary for variables predicting depression.
Predictor	Adj r2	R2	f	p-Value
Function (WOMAC)	0.24	0.24	22.9	<0.001
Age	0.29	0.05	15.0	<0.001
Adj = adjusted, WOMAC = Western Ontario and MacMaster Universities (Osteoarthritis Index).

We report the following findings from this investigation. First, a higher percentage of patients screened positively for depression than has been reported by others. Second, depression scores were related to perceived physical function and not as strongly to measured physical performance tasks. Lastly, lower function and younger age accounted for 29 percent of the variance in depressive symptoms.

Forty-two percent of our sample screened positively for depression with the CES-D. The mean value for the CES-D was 15, compared with a score of 10 reported by other investigators [21-22]. Because our sample was primarily men, and depression is more commonly reported by women [23], this was an unexpected finding. Our sample did report higher levels of pain on the WOMAC pain subscale, which may account for higher reports of depression. Additionally, our sample was drawn from a primarily veteran population. It is well documented that veterans, in general, suffer from symptoms of depression more than nonveterans [24], and this may account for a higher incidence of depression in this sample.

Higher levels of depression in our sample were associated with reported decreased function, more knee stiffness, and more walking impairment than the nondepressed patients. Patients with depressive symptoms, even without a diagnosis of depressive disorder, have poor functioning and these symptoms become clinically meaningful. Thus, persons with knee OA may be more likely to report symptoms of depression and anxiety [25-26]. Sharma et al. demonstrated that the status of mental health (SF-36) predicts physical function (WOMAC), because good mental health is a protector against poor function [27]. Creamer et al. did not find significant correlations between depression (CES-D) and disability scores (WOMAC) [28]. However, other investigators have reported moderately strong relationships between depression and function measured by the Sickness Impact Profile [29] and the Arthritis Impact Scale [30].

Interestingly, in both the nondepressed and depressed patients, moderate relationships were identified between the 6-minute walk distance and the number of stairs climbed and descended, as well as between the WIQ subscales and the SF-36 Physical Composite scores. Moderate relationships were also identified in the nondepressed patients for the WOMAC subscales; these relationships were not present, however, in the depressed patients. This unexpected finding is important because the WOMAC is commonly used in patients with knee OA. We postulate that in depressed patients, function may be so heavily influenced by depressive symptoms that the relationship between actual performance and ability may be lost. Alternatively, one could postulate that the questionnaires and performance indices are measuring different constructs. When examining specific items on the WOMAC function subscale, one can surmise that depressed patients may have difficulty going shopping, getting dressed, or performing domestic duties, and these difficulties may have little to do with their physical ability to carry out these tasks. If this is the case, researchers and clinicians may need to consider the role of depression in self-reported physical function outcomes. In patients who are depressed, an instrument targeted at specific activities, such as the WIQ, may more accurately reflect the patient's physical ability.

Our regression model showed that physical function and age were associated with depression. Previous researchers reported that depression predicts self-reported physical function [3-4]; in our sample, self-reported poor physical function was a predictor of depression.

One limitation of our study may be that this cohort is predominantly male; the reported literature indicated that the severity and distribution of depression are different in males and females [17]. Therefore, analysis results about depression may not be generalized to females. Likewise, these results cannot determine the causal influence of depressive symptoms.

These data indicate that depressive symptoms in overweight and obese patients with OA of the knee are related to their self-reported physical function. Additionally, perceived function and younger age are determinants of depressive symptoms in overweight and obese persons with OA of the knee. More research is needed to determine whether our findings regarding the relationship between depression and perceived function and measured physical performance persist in larger, more diverse samples of patients with OA of the knee.

Go to TOP

Go to the Table of Contents of Vol. 46 No. 9

Last Reviewed or Updated  Wednesday, February 17, 2010 9:25 AM

Current Issue 49-3

Table of Contents
[PDF]

At a Glance

Browse JRRD

VA Home |  Privacy Policy |  FOIA |  Web Policies |  No FEAR Act Data |  Site Index |  USA.gov |  White House |  National Resource Directory |  Inspector General

U.S. Department of Veterans Affairs - 810 Vermont Avenue, NW - Washington, DC 20420