Can Physical Activity Attenuate Aging-related Weight Loss in Older People? The Yale Health and Aging Study, 1982-1994

1
Can Physical Activity Attenuate Aging-related
Weight Loss in Older People? The Yale Health and
Aging Study, 1982-1994

• James Dziura, Carlos Mendes de Leon, Stanislav
  Kasl, and Loretta DiPietro

2
Background

• Body weight increases through middle-age, but
  starts to decline during older age
• This decline occurs primarily as a result of loss
  in lean tissue (sarcopenea)
• Sarcopenea is a primary condition for loss of
  physical and metabolic reserve in aging
• Disuse has been identified as a significant
  factor in lean tissue loss, since changes in body
  composition are mirrored by a gradual decline in
  physical activity in older age.

3
Problem

• Longitudinal studies of physical activity and
  weight change have not been performed in older
  populations, among whom weight loss is a more
  important problem than weight gain and among whom
  the relation between physical activity and weight
  loss is influenced heavily by chronic disease and
  loss of mobility.

4
Purpose
The purpose of this analysis was to determine,
longitudinally, the trajectory of weight loss
with age and the relation between physical
activity and body weight change in a cohort of
people aged 65 years and older.
5
Hypotheses
We hypothesized that 1) body weight would
decline over the follow-up and 2) physical
activity would play an important role in
minimizing weight loss over time.
6
Methods
Subjects Older residents (65 y) of New Haven,
CT enrolled in the Yale Health and Aging Project
one of the funded sites of the Established
Populations for the Epidemiologic Study of the
Elderly (EPESE).




7
Methods

• Subjects
• The New Haven EPESE cohort was a cluster sample
  stratified by three different housing types for
  older people 1) public housing 2) private
  housing and 3) housing in the general community.
• Of the 3,337 eligible residents contacted, 2,812
  (82) enrolled in the study at baseline in 1982.

8
Exclusion Hierarchy
Interested subjects (N2,812)
9
Methods

• Body weight
• Current body weight was self-reported at each
  yearly interview from 1982 to 1990 and then again
  in 1994.
• Each subject could have as many as 10 weight
  assessments over the 12-year follow-up.

10
Methods

• Physical Activity
• Respondents were asked to report the frequency of
  participation within the past month for four
  general types of activities 1) walking 2)
  gardening/housework 3) physical exercise and 4)
  active sports or swimming.
• Responses of never (coded as 0), sometimes (coded
  as 1) or often (coded as 2) were summed over the
  four activities for each respondent to create a
  total activity score (TAS) that ranged from 0 to
  8, with higher scores indicating higher levels of
  baseline activity.

11
Covariables at Baseline

• Age (date of initial interview - reported birth
  date)
• Height (inches)
• Race (white/non-white)
• Education (number of years)
• Chronic conditions (reported number of
  physician-diagnoses of heart disease, cancer,
  stroke, diabetes, liver disease, or
  hypertension)
• Smoking (smoker/non-smoker)
• Functional Ability Score (ADLs, IADLs)
• Mobility Score (reported ability (yes/no) to
  climb stairs, do house work, and walk ½ mile)

12
Statistical Analysis

• Univariate statistics (?SD )
• Multivariable random effects modeling (repeated
  body weight assessments over a maximum of 10
  times (1982-1994) regressed on physical activity,
  while adjusting for covariables). This modeling
  accounts for between-subject variation in
  baseline body weight (intercept) and change in
  body weight with aging (slope as indicated by
  follow-up time variable).

13
Table 1. Cross-sectional distribution of mean
(sd) body weight at each year of follow-up among
those who survived and those who did not survive
to the next interview year.
Survived to next interview year
Did not survive to next interview year
YEAR N Body weight (lbs) N Body weight (lbs)
1982 2199 15431 60 15328
1983 2142 15230 36 15126
1984 1965 15131 79 14829
1985 1832 15232 14 16021
1986 1749 15031 30 15227
1987 1572 14931 82 14426
1988 1397 14931 18 15424
1989 1326 14631 64 15131
1990 1093 14531 40 15324
14
Table 2. Baseline characteristics of the study
sample by survival status over the follow-up,
1982-1994.
Variable Survivors (N905) Non-survivors (N1395)
Age (yrs) 715 767
Sex ( Female) 69 51
Race ( White) 78 81
Education (yrs) 104 94
Disability score 0.130.60 0.431.30
Mobility score 0.39 0.76 0.871.06
Chronic conditions 10.77 10.92
Smokers () 18 23
Height (in) 64.33.58 64.93.71
Weight (lbs) 15530 15332
Total Physical Activity Score 2.621.78 2.112.19
plt0.001 based on t-test for independent
samples plt0.01 based on chi-square test.
15
Figure 2. - Predicted weight loss (95 confidence
bands) over the first 8 years based on random
effects modeling 1982-1990.
Body weight (lbs)
Survivors
Non-survivors
YEAR
Model adjusted for age, sex, housing strata,
and interaction of each with time.
16
Table 3. - Simple effects of physical activity on
weight loss 1982-1994.
Non-survivors (n1,395)
Combined (N2,300)
Survivors (n905)
Variable Estimate 95 CI Estimate 95 CI Estimate 95 CI
Total Activity Score Total Activity Score Time 1.29 0.16 0.57, 2.00 0.10, 0.22 1.19 0.05 0.13, 2.25 -0.01, 0.11 1.19 0.20 0.21, 2.17 0.09, 0.31
Based on random effects modeling adjusted for
time and time2. Regression estimates for
TASTime interaction represent the change in the
rate (per year) of weight loss per unit change in
the study variable.
17
Table 4. Parameter Estimates from the Random
Effects Model Determining the Effects of Physical
Activity on Aging-Related Weight Change.
Survivors (n905)
Non-survivors (n1,395)
Combined (N2,300)
Variable Parameter estimate 95 CI Parameter estimate 95 CI Parameter estimate 95 CI
Baseline Age -1.04 -1.21, -0.87 -0.78 -1.12, -0.45 -1.12 -1.34, -0.91
Time -1.77 -2.19, -1.35 -0.25 -0.73, 0.24 -2.31 -3.14, -1.47
Time2 -0.04 -0.05, -0.02 -0.04 -0.06, -0.03 -0.13 -0.16, -0.09
Age Time -0.07 -0.08, -0.05 -0.04 -0.06, -0.02 -0.03 -0.06, -0.01
Chronic condition 3.94 1.73, 6.16 9.97 3.97,16.0 2.38 -2.63, 7.38
Chronic condition Time -0.15 -0.36, 0.06 -0.34 -0.75, -0.01 0.08 -0.56, 0.72
TAS -0.83 -1.76, 0.10 0.25 -1.43, 1.92 -1.03 -2.93, 0.87
TAS Time 0.09 0.02, 0.15 -0.03 -0.12, 0.06 0.18 0.00, 0.37
TASChronic condition ----- ----------- -0.91 -2.76,0.95 -0.37 -2.18, 1.43
TASChronic conditionTime ----- ---------- 0.12 0.01,0.24 -0.10 -0.32, 0.12
Mobility -3.41 -6.08, -0.74 -6.12 -11.2, -1.08 -3.12 -6.38, 0.14
MobilityTime 0.36 0.09, 0.63 -0.22 -0.53, 0.08 0.41 -0.02, 0.84
Smoking Smoke Time -10.30 0.24 -12.9, -7.70 -0.02, 0.50 -10.20 0.23 -14.5, -5.98 -0.04, 0.49 -10.05 0.52 -13.4, -6.70 0.07, 0.97
also adjusted for sex, race, education, height,
functional disability score, and housing.
18
Fig. 3 - Parameter Estimates (95 CI) from the
Fully-Adjusted Random Effects Model Determining
the Effects of Physical Activity Score on
Attenuation of Weight Loss by Survival and
Chronic Disease Status.
Attenuation of weight loss (lbs/year)
0.15 (0.00, 0.30)
0.06 (-0.12, 0.25)
0.09 (0.02, 0.15)
0.09 (0.00, 0.17)
-0.02 (-0.11, 0.07)
Combined (N2300)
Survivors with no chronic disease (n350)
Survivors with chronic disease (n555)
Non-survivors with no chronic disease (n422)
Non-survivors with chronic disease (n973)
19
Summary

• The oldest people started off the follow-up
  period weighing less than their younger
  counterparts
• The magnitude of the inverse relation between age
  and body weight was greater among nonsurvivors
  compared with survivors

20
Summary

• We observed a curvilinear rate of weight loss
  with aging that varied by survival status and by
  chronic disease.
• The ability of physical activity to attenuate
  aging-related weight loss was observed only among
  those with chronic disease, and was most
  pronounced among those with chronic disease who
  did not survive the follow-up period.

21
Conclusion
If we were to consider weight loss a somatic
indicator of frailty in older people, it appears
as though even among older people afflicted with
chronic conditions that contribute to frailty, an
active lifestyle can attenuate the rate of
failure, thereby maintaining some quality of
health and function in older age.