Association of Handgrip Strength and Muscle Mass with Dependency in (Instrumental) Activities of Daily Living in Hospitalized Older Adults -The EMPOWER Study

The ‘Evaluation of Muscle parameters in a Prospective cohort of Older patients at clinical Wards Exploring Relations with bed rest and malnutrition’ (EMPOWER) study is an observational, prospective inception cohort study. A total of 838 subsequent hospitalized patients were screened for eligibility for inclusion from April 2015 to December 2015. Patients were included if they were aged 70 years and older and admitted to one of four clinical wards (acute admission, internal medicine, neurosurgery and orthopedics, traumatology) of the VU University Medical Center, Amsterdam, The Netherlands, a middle-sized teaching hospital. Patients were excluded if they were terminally ill, had an expected length of stay in the hospital shorter than 24 hours, were admitted at another hospital and relocated to the VU University Medical Center, were nursed in isolation rooms or were not able to understand the Dutch language. The EMPOWER study included a total of 378 patients. Patients were assessed within 48 hours after hospital admission, comprising assessment of patient characteristics, measurements of muscle strength and muscle mass, and ADL and IADL dependency. ADL and IADL dependency was re-assessed at three months after discharge using telephone calls as described earlier (14). This study was reviewed and approved by the medical ethical committee of the VU University Medical Center, Amsterdam, The Netherlands. Informed consent was obtained from all included patients.

At admission, the following items were assessed during an interview by a nurse: living situation, use of walking aid, risk of malnutrition defined by the Short Nutritional Assessment Questionnaire (SNAQ) score using a score ≥2 out of 7 points (15), risk of delirium (16), cognitive impairment defined by the Six Item Cognitive Impairment Test (6CIT) score using a score ≥11 out of 28 points (17).

The following items were collected from medical records: age, sex, multimorbidity, polypharmacy, acute or elective admission and length of stay. Multimorbidity was defined as two or more of the following diseases: cerebrovascular accident/transient ischemic attack, chronic obstructive pulmonary disease, hypertension, heart failure, chronic kidney disease, diabetes mellitus, malignancy, peripheral vascular disease, thyroid disease, Parkinson’s disease, arthrosis, rheumatoid arthritis and osteoporosis. Polypharmacy was defined as five or more medications (18).

Weight was measured to the nearest 0.1 kilograms using a weighing chair. Self-reported weight was used when measurement of weight was impossible. Knee height was measured to the nearest 0.1 centimeter to estimate actual height to the nearest 1 cm using the Longitudinal Aging Study Amsterdam (LASA) formula (19). BMI was calculated to the nearest 0.1 kilogram/meter2 as weight divided by height squared.

HGS was measured using a hydraulic hand dynamometer (Jamar, Sammons Preston Rolyan, IL, USA). Measurements were performed according to instructions recommended by the American Society of Hand Therapists (20), i.e. patients were asked to sit straight up in a chair or on the edge of the bed, flex their elbow about 90 degrees, not support their elbows and squeeze maximally. If patients were unable to sit, patients were asked to put the bed in an angle of approximately 30 degrees, flex their elbow about 90 degrees, not support their elbows and squeeze maximally under verbal encouragement. Two measurements per hand side alternately were performed. The maximal value of all four measurements was used for analysis.

Body composition was measured using direct segmental multi-frequency bioelectrical impedance analysis (DSM-BIA) (InBody S10, Biospace Co., Ltd, Seoul), which is a validated and standardized method of estimating skeletal muscle mass compared to DXA (21). Patients were asked to lie down in bed, spread their arms to about a 15 degrees’ angle, arms not touching the trunk and to spread legs to shoulder width, thighs not touching each other. If patients were not able to lie down flat, measurements were performed in maximum supine position that could be reached. BIA measurements were not performed when the patient had a pacemaker or other implanted electronic device (n = 25) and when electrodes could not be placed on fingers or ankles (n = 21). The following measures of muscle mass were used for analyses: skeletal muscle mass (SMM) in kilograms, SMM as a percentage of total body weight (SMM%) (8) and skeletal mass index as SMM divided by height squared (SMI) (22). Fat mass in kilograms was obtained during measurement of body composition and was measured to the nearest 0.1 kilograms.

ADL dependency was assessed using the Katz score (23) ranging from 0 to 6 points, a lower score meaning greater dependency. For the cross-sectional analysis, ADL score was dichotomized: a score of <6 points was considered dependent and a score of 6 points was considered independent (24). IADL dependency was determined using the Lawton and Brody score (25) ranging from 0 to 8 points, a lower score meaning greater dependency. IADL score was dichotomized: a score of <8 points was considered dependent and a score of 8 points was considered independent (26). Due to a protocol amendment at a later stage in which the assessment of IADL was added to the protocol, IADL data at admission are available in sample of 262 (69.3%) patients.

For longitudinal analysis, the change in ADL and IADL was defined as the score at follow-up assessment minus the score at admission (change in ADL and IADL); positive change ADL and IADL scores indicating a decrease in dependency and negative change in ADL or IADL scores indicating an increase in dependency.

Continuous variables were presented as mean with standard deviation (SD) if normally distributed and as median with interquartile range (IQR) if not normally distributed.

Cross-sectional analysis of the associations of HGS, SMM, SMM% and SMI (independent variables) with ADL and IADL dependency (dependent variables) at admission were analyzed using binary logistic regression analysis (independency = 1, dependency = 0). In model 1, analyses were adjusted for age. In model 2, analyses were additionally adjusted for multimorbidity and weight (kg) and height (m) (for HGS) or fat mass (kg) (for SMM and SMI) or weight (kg) (for SMM%) (27).

Longitudinal analyses of the association of HGS, SMM, SMM% and SMI with the change in ADL and IADL scores were analyzed using linear regression analyses. In model 1, analyses were adjusted for age and ADL or IADL score at admission. In model 2, analyses were additionally adjusted for multimorbidity and weight (kg) and height (m) (for HGS) or fat mass (kg) (for SMM and SMI) or weight (kg) (for SMM%) (27).

Statistical analyses were performed using the Statistical Package for the Social Sciences, version 22 (IBM SPSS Statistics, IBM Corporation, Chicago, IL). P-values lower than 0.05 were considered statistically significant. P-values lower than 0.1 were considered a trend.