Body Mass Index and Physical Function in Older Women Caroline M. Apovian,* Carolin M. Frey,† G. Craig Wood,‡ Joanne Z. Rogers,§ Christopher D. Still,§ and Gordon L. Jensen¶
Abstract APOVIAN, CAROLINE M., CAROLIN M. FREY, G. CRAIG WOOD, JOANNE Z. ROGERS, CHRISTOPHER D. STILL, AND GORDON L. JENSEN. Body mass index and physical function in older women. Obes Res. 2002;10:740 –747. Objective: We modified existing standardized measurement tools in the Physical Performance Test and tasks from the Frailty and Injuries: Cooperative Studies of Intervention Technique Study to evaluate physical function in older women. Our objectives were (1) to characterize physical function themes based on combinations of tasks (deriving factors or components) and (2) to quantify the correlation between derived factors and body mass index (BMI). Research Methods and Procedures: Nutrition risk screens from enrollees in a Medicare-managed risk program served as the sampling frame. To obtain adequate representation for a range of BMI, a random sample was obtained of 90 women from the following BMI strata: BMI, 22 to ⬍27 kg/m2; BMI, 27 to ⬍30 kg/m2; and BMI, ⱖ30 kg/m2. Subjects were asked to perform a series of 18 functional tasks during a home visit. Results: The mean age was similar in the three BMI groups with an overall mean age of 71 ⫾ 4.9 years (SD). Factors characterized by lower-body function, upper-body function, coordination, and strength were responsible for 30%, 11%, 9%, and 9% of the variance in task scores, respectively. BMI, controlling for age, explained 5%, 14%, 3%, and 0%
Received for review July 30, 2001. Accepted for publication in final form April 29, 2002. *Department of Endocrinology, Diabetes, and Nutrition, Boston University School of Medicine, Boston, Massachusetts; †Department of Health Services and Community Research, Children’s National Medical Center, Washington, DC; ‡Clinical Research Department and §Department of Nutrition, Geisinger Medical Center, Danville, Pennsylvania; and ¶Vanderbilt Center for Human Nutrition, Nashville, Tennessee. Address correspondence to Caroline M. Apovian, M.D., Director, Nutrition and Weight Management Center, Associate Professor of Medicine, Boston University School of Medicine, Building D, Suite 614, 88 East Newton Street, Boston, MA 02118-2393. E-mail:
[email protected] Copyright © 2002 NAASO
740
OBESITY RESEARCH Vol. 10 No. 8 August 2002
of the variation in these factors, respectively. Higher BMI is associated significantly with poorer upper- and lower-body function but is not associated significantly to strength or coordination. Discussion: Higher BMI seems to differentially impede specific aspects of physical function, especially upper-body function, and to a lesser extent, lower-body function. BMI does not seem to be associated with levels of coordination or strength. Better understanding of how BMI impacts physical function will aid in the design of interventions to promote independent living in elderly, obese women. Key words: physical function, elderly, women
Introduction Several studies have correlated excess body weight with the development of disability in older persons. High body mass index (BMI), a measure of weight standardized for height (kilograms per square meter), is positively associated with present disability (1,2) and with risk of developing impaired physical function among older men and women (3,4). Other studies showed an increased risk for disability in association with chronic disease, including cardiovascular disease, arthritis, diabetes, and pulmonary disease (2,5– 8). Recent prospective data suggest that BMI is a strong predictor of long-term risk for mobility disability in older women, and that this risk persists even to very old age (9). In contrast, BMI below the 75th percentile (25.4 kg/m2 for men and 25.2 kg/m2 for women) was associated with a high likelihood for continued physical ability and high level of physical function (10). The mechanisms through which body mass may affect disability have not been identified, although it is suspected that part of the increased risk of disability of overweight persons is caused by the development of chronic disease related to obesity, particularly cardiovascular disease and arthritis (11). However, other factors that are likely to be contributory to disability, especially in the elderly obese, include diminished exercise tolerance, frailty, and social or
BMI and Physical Function in Older Women, Apovian et al.
psychological disadvantages (12). Obesity may contribute to disability by decreasing endurance, by increasing the energy demands of ventilation, and by altering pulmonary function (13–16). Obesity may also impede mobility and flexibility (17). This suggests that the potential exists to design interventions that directly target obesity-related disability in the elderly, separate from treatments for obesityrelated chronic disease. Unfortunately, current research methodology relies on measures of disability for the elderly (18 –24) that do not capture the specific effect of excess body weight on physical function. The Physical Performance Test (PPT; see Table 1) is an objective, quantifiable test assessing multiple domains of physical function using observed performance of tasks that simulate activities of daily living (24). It can be completed in ⬍10 minutes using only a few simple props and can, therefore, be readily given in the office or community setting. This test demonstrates better sensitivity in detecting disabilities associated with chronic diseases affecting physical functioning than conventional self-reported functional scales. In general, timed physical performance measures have been found to be strong predictors of future functional dependence among nondisabled, free-living, older persons (25). A preliminary study tested the validity of the PPT as a predictor of disability specifically caused by obesity (26). The summary PPT score was associated significantly with measures of body fat. However, analysis of the individual tasks in the PPT did not clearly identify which tasks contributed most to this overall association. This suggested that additional tasks might enhance the yield of information from physical performance testing and that combinations of tasks may be more informative than single tasks in describing obesity-related decrements in performance. The Frailty and Injuries: Cooperative Studies of Intervention Techniques (FICSIT) trials were the first nationally sponsored set of clinical trials concerning physical frailty and risk for injuries in later life (27). The results are intended to serve as a reference in designing health interventions for older persons. A modification of the PPT combined with some of the tests used in FICSIT were used to evaluate physical function in elderly women over a wide range of BMI.
Research Methods and Procedures Subjects for this study were gathered by random sampling through a nutrition screening program of the Geisinger Health Plan that is administered to all enrollees in a Medicare Managed-Risk Program (28). More than 100 Geisinger Clinic sites throughout largely rural central and eastern Pennsylvania implemented a one-page level II Screen developed by the Nutrition Screening Initiative (29). Clinic staff recorded height and weight, and enrollees completed remaining items during visits for usual care. Clinical staff forwarded completed assessment forms to our research center. A possible problem in ensuring equal representation of
women across the range of BMI in this study population involved the potential for higher rates of refusal among women with higher BMI; therefore, a stratified random sampling scheme was used. Women were stratified into three groups defined by tertiles of healthy to high BMI: 22 to ⬍27 kg/m2; 27 to ⬍30 kg/m2; and ⱖ30 kg/m2. Women with BMI ⱕ22 kg/m2 were excluded because of the likelihood of disability secondary to frailty and sarcopenia. Women within each stratum, who were 65 years of age or older, were randomly selected as potential subjects. Medical record review excluded potential subjects who were not community-dwelling; had a history of severe depression or cognitive disability that could compromise meaningful consent and participation; had any disease or event that might compromise physical function independent of obesity including rheumatoid arthritis, Paget’s disease of bone, or other bone diseases; or had disability caused by prior trauma. Potential subjects with diabetic neuropathy, osteoarthritis, sleep apnea, hypertension, hyperlipidemia, or any other comorbidity of obesity were eligible for the study. The study coordinator contacted eligible women by phone to invite study participation. The Geisinger Medical Center Institutional Research Review Board approved this study. Study staff obtained written informed consent in person from each subject at the home visit. A “mini-mental status” exam (30), given to each subject at the home visit, excluded two women who scored less than the average for the population reference group based on age and educational level, even though prior medical record review did not identify cognitive deficits (one woman with three errors and a high-school education, one woman with four errors and an elementary-school education). Study staff measured height and weight to verify qualifying BMI and to correctly classify women into BMI strata. This yielded several reclassifications between initial BMI strata. Study staff then administered a composite set of functional performance tests to study subjects. Focusing on mobility and endurance, a composite set of tasks was designed to combine 8 of the 9 PPT tasks (excluding a four-flight stair climb) with 10 tasks used in the FICSIT trials (27). The added tasks included an accelerated chair stand, hand-grip dynamometry (dominant and nondominant), static balance (feet together, foot halfway in front of the other, and foot totally in front of the other), one-legged stance (right and left), and one-legged stance with eyes closed (right and left). All timed measures were determined with a stopwatch. Two of the measurements—a 50-foot walk and climbing one flight of stairs—were revised from the description in Reuben and Siu (24) to accommodate testing in the home environment. The 50-foot walk was modified for home use by retracing steps after walking 25 feet. Climbing one flight of stairs (12 steps) was accomplished using a stair stepper built specifically for this purpose. Quartile rankings, based on the full complement of OBESITY RESEARCH Vol. 10 No. 8 August 2002
741
BMI and Physical Function in Older Women, Apovian et al.
Table 1. Scoring of functional tasks Task
Recorded value
Write sentence*
Time (seconds)
Simulate eating*
Time (seconds)
Lift book*
Time (seconds)
Put on and remove a jacket*
Time (seconds)
Pick up penny from floor*
Time (seconds)
Circle turn*
Continuous (yes/no) and steady (yes/no)
Walk 50 feet†‡
Time (seconds)
Stair climbing†‡
Time (seconds)
Accelerated chair stand§
See right
Hand-grip dynamometry—dominant hand†§
Average force of three attempts (kg)
742
OBESITY RESEARCH Vol. 10 No. 8 August 2002
Task score 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 2 2 4 1 2 3 4 0 1 2 3 4 1 2 3 4 0 1 2 3 4
if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if if
unable to do ⱖ20.5 seconds 15.5 to 20.0 seconds 10.5 to 15.0 seconds ⱕ10.0 seconds unable to do ⱖ20.5 seconds 15.5 to 20.0 seconds 10.5 to 15.0 seconds ⱕ10.0 seconds unable to do ⱖ6.5 seconds 4.5 to 6.0 seconds 2.5 to 4.0 seconds ⱕ2.0 seconds unable to do ⱖ20.5 seconds 15.5 to 20.0 seconds 10.5 to 15.0 seconds ⱕ10.0 seconds unable to do ⱖ6.5 seconds 4.5 to 6.0 seconds 2.5 to 4.0 seconds ⱕ2.0 seconds discontinuous and unsteady discontinuous and steady continuous and unsteady continuous and steady ⱖ20.5 seconds 17.5 to 20.0 seconds 15.5 to 17.0 seconds ⱕ15.0 seconds unable to do ⱖ38.5 seconds 29.5 to 38.0 seconds 25.5 to 29.0 seconds ⱕ25.0 seconds unable to do uses assistance device to push up uses arm to push up able to stand up with arms crossed unable to do ⱕ21.2 kg 21.2 ⬍ force ⱕ 25.2 kg 25.2 ⬍ force ⱕ 28.5 kg ⬎28.5 kg
BMI and Physical Function in Older Women, Apovian et al.
Table 1. continued Task
Recorded value
Hand-grip dynamometry—non-dominant hand†§
Average force of three attempts (kg)
Static balance—foot halfway in front of the other§ Static balance—foot totally in front of the other§ One-legged stance—eyes open—right side§ One-legged stance—eyes open—left side§ One-legged stance—eyes closed—right side§ One-legged stance—eyes closed—left side§
Time Time Time Time Time Time
up up up up up up
to to to to to to
10 10 10 10 10 10
seconds seconds seconds seconds seconds seconds
Task score 0 if unable to do 1 if ⱕ19.3 kg 2 if 19.3 ⬍ force ⱕ 23.0 kg 3 if 23.0 ⬍ force ⱕ 26.1 kg 4 if ⬎26.1 kg Actual time Actual time Actual time Actual time Actual time Actual time
* Unmodified task, timed tasks were rounded to the nearest half second prior to scoring. † Score derived as quartile of non-missing recorded value for completed task with zero assigned for those tasks the subject was unable to do. ‡ Modified Physical Performance Test task, timed tasks were rounded to the nearest half second prior to scoring. § Frailty and Injuries: Cooperative Studies of Intervention Techniques task.
study data, replaced actual times for the 50-foot walk, stair climbing, and dominant and non-dominant hand-grip dynamometry. Those subjects unable to complete the task received scores of zero. This scheme handled uncompleted FICSIT tasks in a manner comparable with PPT tasks. Table 1 reports the correspondence between recorded values and scores for each of the 18 tasks. Although it is generally preferable to use continuous valued measurements such as those involving time where possible, it was necessary to use ordinal assignments to allow for the inability to complete a task (equivalently, an infinite time). The order of scores is consistent across tasks such that lower numbers indicate difficulty in completing tasks or poor performance, and higher numbers indicate successful completion or excellent performance. For some PPT tasks, this involved reversing the order of task scores. Such reordering does not fundamentally change the properties of the measurement. Medians and bootstrapped 95% confidence intervals summarized task performance for each BMI group (Table 2). Medians permit meaningful summarizing of timed values that were right-skewed. Factor analysis (31) identified a parsimonious set of thematically meaningful factors comprised by weighted combinations of tasks. The sample size of 90 subjects was chosen because of a rule of thumb that suggests that a minimum of five subjects per variable—18 functional tasks here— be used to support a factor analysis. However, some data reduction before factor analysis seemed to be prudent after initial inspection of the data. This improves the statistical power for this analysis. The factor analysis excludes scores for the “Static Balance-Feet To-
gether” task, which was the maximum of 10 seconds for all subjects. Further data reduction was achieved by averaging paired (left/right or dominant/non-dominant) measures. This left 14 variables for input into the factor analysis. A latent root criterion guided the retention of factors with a latent root of at least one. Varimax rotation of retained factors estimated factor loadings for the purpose of interpretation. Multiple linear regression analysis quantified associations between BMI, while controlling for age, with derived components characterized by specific dimensions of physical function. Data analyses used SAS software (Statistical Analysis Systems, Cary, NC).
Results
A total of 90 women, with a BMI ⬎22 kg/m2, were enrolled in this study, with two exclusions after the minimental status exam. The 88 remaining participants were assigned BMI categories; 32 women were in the lowest BMI group, 27 women were in the middle BMI group, and 29 women were in the highest. There was no difference in age between the three BMI groups (p ⫽ 0.276; 71 ⫾ 5 years). A comparison of women who were in the sampling frame but who did not participate in the study (n ⫽ 951) to those enrolled (n ⫽ 88) partially addressed the representativeness of the study sample. The study and nonstudy groups did not differ with respect to BMI (29.1 and 29.7, respectively; p ⫽ 0.297), and age (70.8 and 70.3 years; respectively, p ⫽ 0.412). Additionally, the two groups did not differ with OBESITY RESEARCH Vol. 10 No. 8 August 2002
743
BMI and Physical Function in Older Women, Apovian et al.
Table 2. Medians and bootstrap 95% confidence intervals (confidence intervals shown in parentheses) for selected* functional tasks and standardized factor scores Tasks
BMI 22 to
