Full text of DOI:10.1093/gerona/56.12.B497

Journal of Gerontology: BIOLOGICAL SCIENCES
Copyright 2001 by The Gerontological Society of America
2
001, Vol. 56A, No. 12, B497–B502
Familial Aggregation of 7-Year Changes in
Musculoskeletal Fitness
1
1
2
3
Peter T. Katzmarzyk, Norman Gledhill, Louis Pérusse, and Claude Bouchard
1School of Kinesiology and Health Science, York University, Toronto, Ontario, Canada.
2Division of Kinesiology, Department of Preventive Medicine, Laval University, Ste-Foy, Quebec, Canada.
3Pennington Biomedical Research Center, Louisiana State University, Baton Rouge.
The purpose of this study was to determine the degree of familial resemblance in baseline and
7
-year changes in musculoskeletal fitness. Data from the 1981 Canada Fitness Survey and the
Campbell’s Survey 7-year follow-up were used. The sample consisted of 1264 people (635
males and 629 females) between the ages of 7 and 69 years for whom measurements of muscu-
loskeletal fitness were available at baseline. A subsample of 834 people had measurements at
both baseline and 7-year follow-up. Sit-and-reach trunk flexibility, number of push-ups without
time limit, number of sit-ups in 60 seconds, and hand-grip strength were used as indicators of
musculoskeletal fitness. The data were adjusted for the effects of age and body mass index (and
baseline level of the variable for changes) by using regression procedures, and they were stan-
dardized to zero mean and unit variance within each of the four sex-by-generation groups (fa-
thers, mothers, sons, and daughters). Familial correlation models were fitted to the data by using
the computer software SEGPATH. The results indicate significant familial resemblance for all
indicators of musculoskeletal fitness for baseline measures and 7-year changes. The heritabili-
ties, or the percentages of the total variance attributable to heredity, were 64% for trunk flexibil-
ity, 37% for push-ups, 59% for sit-ups, and 48% for grip strength. Similarly, heritabilities for the
change scores were 48% for trunk flexibility, 52% for push-ups, 41% for sit-ups, and 32% for
grip strength. The results suggest that familial, and perhaps genetic, factors are important in ex-
plaining the variance in musculoskeletal fitness not only cross-sectionally but also for changes
over time.
USCULOSKELETAL fitness refers to those compo-
nents of fitness that are related to flexibility, muscular
present study (grip strength, push-ups, sit-ups, and sit-and-
reach trunk flexibility) are influenced by genetic factors (8).
Family studies (9–14), studies of twins (15–19), and mixed
designs (20) have all shown that indicators of musculoskel-
etal fitness are inherited characteristics, although estimates
of heritability vary from measurement to measurement.
The available evidence for a genetic component to mus-
culoskeletal fitness comes from cross-sectional studies of
families and twins. We know of no published studies related
to the heritability of changes in musculoskeletal fitness over
time. Given the dearth of information regarding the genetics
of age-related changes in musculoskeletal fitness, the pur-
pose of this study was to determine the degree of familial
resemblance in baseline and 7-year changes in musculoskel-
etal fitness. The 1988 Campbell’s Survey database, which
was a longitudinal follow-up of the 1981 Canada Fitness
Survey, was used to address the aims of the study.
M
strength, muscular power, and muscular endurance. Muscu-
loskeletal fitness is positively related to health status across
the life span (1,2) and to functional ability in the elderly
population (3,4). In general, levels of musculoskeletal fit-
ness tend to decline over the life span (5,6); however, there
is considerable individual variability in changes over time
5). Thus, it is important to understand the determinants of
musculoskeletal fitness as well as the determinants of
changes that occur with aging.
(
Most phenotypes, or traits, are influenced by both genetic
and environmental factors to varying degrees. The total phe-
notypic variability (V ) in a given trait can be modeled as
P
having genetic (V ), environmental (V ), and genotype–
G
E
environment interaction (VG ꢀ E) components, in addition (7)
to unmeasured residual error (e): V ꢁ V ꢂ V ꢂ V
ꢂ
G ꢀ E
P
G
E
e. The heritability of a trait is defined as the proportion of the
total phenotypic variance that can be attributed to genetic
METHODS
2
factors (i.e., h ꢁ V /V ), and estimates are typically derived
G
P
from studies of nuclear families, monozygotic (MZ) and
dizygotic (DZ) twin pairs, and extended family pedigrees.
The results of several studies have suggested that indica-
tors of musculoskeletal fitness, such as those used in the
Sample
The Campbell’s Survey was a 7-year longitudinal follow-
up to the 1981 Canada Fitness Survey (CFS) and contains
information on 4345 individuals ranging in age from 7 to 69
B497

B498
KATZMARZYK ET AL.
years. The original CFS sample was selected by Statistics
Canada to be representative of the Canadian population and
contains information on individuals from urban and rural ar-
eas of every province (21). Data collection revolved around
households, and the first person to be contacted by the CFS
team was designated as the reference person. All individuals
in the household were then identified by their relationship to
the reference person. Thus, the family structures of the en-
tire sample could be reconstructed from this information.
The present sample is limited to nuclear families with at
least two biologically related individuals (mothers, fathers,
sons, or daughters), which yielded a total of 1264 people
cedures (23), as explained in detail elsewhere (24). Briefly,
each measure was regressed on BMI and up to a cubic poly-
2
3
nomial in age (age, age , age ) by using forward stepwise
regression (mean regression) retaining terms significant at
the 5% level, within sex-by-generation groups (mothers, fa-
thers, sons, and daughters). The change scores (ꢄ) were fur-
ther adjusted for the effects of ꢄBMI and the baseline level
of the phenotype. The residuals from the mean regressions
were retained and regressed on BMI and up to a cubic poly-
nomial in age (variance regression) in a forward stepwise
manner to test for heteroscedasticity. Heteroscedasticity
was present if any of the predictor variables entered the
variance regression at the 5% level of significance. In the
presence of significant heteroscedasticity, the final pheno-
type was calculated as the residual from the mean regres-
sion divided by the square root of the predicted score from
the variance regression. In the absence of heteroscedastic-
ity, the residual from the mean regression was used as the fi-
nal phenotype. The final phenotypes were standardized to a
mean of zero and unit variance within sex-by-generation
groups (mothers, fathers, sons, and daughters) prior to fur-
ther analysis.
(
635 males and 629 females) for whom measurements of
musculoskeletal fitness in 1981 (baseline) were available in
the Campbell’s Survey database. The sample was distrib-
uted among 502 nuclear families, with an average family
size of 2.75 people. A subsample of 834 people had mea-
surements of musculoskeletal fitness at both baseline and
follow-up, which was used for the longitudinal analyses.
The smaller sample size for the longitudinal analyses re-
sulted because many participants did not complete the mus-
culoskeletal measurements at the second visit; rather they
had only anthropometric or questionnaire data available.
This may introduce some bias into the change scores if it
was those people that decreased in fitness the most that did
not complete the second fitness assessment. Although this
question cannot be answered, there were no significant dif-
ferences between the two groups at baseline.
Familial Correlation Model
As a test of familial aggregation in the measures of mus-
culoskeletal fitness, an analysis of variance (ANOVA) was
used to compare the between-family to within-family vari-
ances, using the family identification number as the depen-
dent variable. Hypotheses regarding familial resemblance in
musculoskeletal fitness were then tested by using the com-
puter program SEGPATH (25). Familial correlation models
were fitted directly to the data under the assumption that the
family data follow a multivariate normal distribution. The
sex-specific correlation model was based on four types of
relatives: fathers (F), mothers (M), sons (S), and daughters
(D), giving rise to eight familial correlations (one spouse,
FM; four parent-offspring, FS, MS, FD, and MD; and three
sibling, SS, DD, SD). A series of nested (reduced) models
were compared with a general model in which all parame-
ters were estimated by using tests of the maximum-likeli-
hood ratio, defined as the difference in minus twice the log
likelihood (ꢅ2 ln L). Asymptotically, the log-likelihood ra-
Measurements
All measurements were made following the standardized
procedures of the CFS (22). Stature and body mass were
measured to the nearest millimeter and 0.1 kg, respectively,
2
and the body mass index (BMI; kg/m ) was calculated.
Hand-grip strength was measured with a Stoelting adjustable
dynamometer (C.H. Stoelting Co., Chicago, IL). Participants
held the dynamometer at the level of the thigh in line with
the forearm and were instructed to squeeze vigorously to ex-
ert maximum force. Maximal grip strengths of two trials for
the left and right hands were summed to provide a single
measure of grip strength (kg). The number of push-ups com-
pleted without time limit (n) and the number of sit-ups per-
formed in 60 seconds (n/min) were used as indicators of
muscular endurance. Participants performed sit-ups from the
supine position, with their fingers behind their ears, their an-
kles held, and their knees flexed 90ꢃ. A complete sit-up re-
quired touching the knees to the elbows. For push-ups, males
balanced from the toes, whereas females balanced from the
knees. Each push-up required a cycle of straightening of the
elbows to the chin touching the floor, with a straight back.
Finally, a sit-and-reach test was used to assess trunk flexibil-
ity. Participants reached toward their toes, with their knees
flat on the floor. The test was repeated twice, with the maxi-
mum value recorded to the nearest 0.5 cm. A trunk flexibility
score of 25 cm is equivalent to touching the floor.
2
tio follows a ꢆ distribution with degrees of freedom equal
to the difference in the number of parameters estimated un-
der the two hypotheses (26). Null hypotheses concerning
the strength of the familial resemblance included no familial
resemblance (FM ꢁ FS ꢁ FD ꢁ MS ꢁ MD ꢁ SD ꢁ SS ꢁ
DD ꢁ 0), no sibling resemblance (SD ꢁ SS ꢁ DD ꢁ 0), no
parent-offspring resemblance (FS ꢁ FD ꢁ MS ꢁ MD ꢁ 0),
and no spousal resemblance (FM ꢁ 0). A series of null hy-
potheses, including no sex differences in offspring (FS ꢁ
FD, MS ꢁ MD, SS ꢁ DD ꢁ SD), no sex differences in off-
spring or parents (FS ꢁ FD ꢁ MS ꢁ MD, SS ꢁ DD ꢁ SD),
and no sex or generation differences (FS ꢁ FD ꢁ MS ꢁ
MD ꢁ SS ꢁ DD ꢁ SD), and all correlations being equal
(
FM ꢁ FS ꢁ FD ꢁ MS ꢁ MD ꢁ SS ꢁ DD ꢁ SD) were
Data Adjustments
also tested.
Baseline values and changes in the musculoskeletal fit-
ness measures were adjusted for the effects of age and BMI
in both the mean and variance by using SAS regression pro-
Akaike’s information criterion (AIC), defined as ꢅ2 ln L
plus twice the number of parameters estimated, was used to
judge the fit of the models (27). The model with the lowest

GENETICS OF MUSCULOSKELETAL FITNESS
B499
Table 1. Descriptive Statistics for Fathers, Mothers, Sons, and
Daughters for Baseline Measures and 7-year Changes (ꢄ)
AIC is the best fitting, or most parsimonious. The best-fit-
ting model for each phenotype was derived by combining
all nonrejected hypotheses, and the combination with the
lowest AIC was selected.
Parameter
n
Mean
SD
n
Mean
SD
Fathers
38.2
25.2
105.0
ꢅ7.9
27.4
ꢅ3.0
17.9
Sons
12.4
18.6
55.8
36.5
27.0
0.4
15.7
7.9
33.8
2.6
Age (y)
368
366
368
224
359
217
353
214
351
214
10.2
3.3
16.9
12.0
9.3
5.2
10.3
8.0
267
266
267
200
267
200
257
193
263
196
5.3
3.1
28.7
22.1
7.6
RESULTS
2
Body mass index (kg/m )
Grip strength (kg)
ꢄ Grip strength (kg)
Trunk flexibility (cm)
ꢄ Trunk flexibility (cm)
Push-ups (n)
ꢄ Push-ups (n)
Sit-ups (n)
ꢄ Sit-ups (n)
The descriptive characteristics of the sample are pre-
sented in Table 1. The mean ages of the fathers and mothers
were 38.2 years and 36.2 years, respectively, and the mean
ages of the sons and daughters were 12.4 years and 12.1
years, respectively. The parents, on average, had decreases
in grip strength, trunk flexibility, push-ups, and sit-ups over
the 7 years, whereas the sons had a mean increase in all
measures. The daughters had mean increases in grip
strength and trunk flexibility, and small decreases in push-
ups and sit-ups. Table 2 presents the results of the data-
7.6
11.1
14.5
11.1
11.4
ꢅ3.8
26.2
ꢅ6.0
9.7
8.3
Mothers
36.2
22.9
61.7
ꢅ5.8
31.2
ꢅ2.3
15.8
ꢅ4.2
19.1
Daughters
12.1
Age (y)
Body mass index (kg/m )
401
401
401
259
391
251
380
242
372
233
9.5
3.6
10.7
9.4
7.9
6.0
10.1
10.2
10.1
8.2
228
228
228
151
228
150
223
148
225
150
5.7
3.7
14.4
14.2
6.8
2
18.7
40.6
15.6
31.3
1.7
19.0
ꢅ1.0
29.1
ꢅ0.6
2
3
adjustment procedures. Age, age , age , and BMI accounted
for up to 82% of the variability in the baseline measures,
Grip strength (kg)
ꢄ Grip strength (kg)
Trunk flexibility (cm)
ꢄ Trunk flexibility (cm)
Push-ups (n)
ꢄ Push-ups (n)
Sit-ups (n)
ꢄ Sit-ups (n)
2
3
whereas age, age , age , ꢄBMI, and baseline values of the
phenotype accounted for between 5.9% and 62.4% of the
variance in the change scores. Heteroscedastic effects were
minor, accounting for between 0.8% and 9.9% of the vari-
ance in 9 of the 32 regressions.
6.3
12.4
13.9
10.3
11.3
ꢅ3.8
The ANOVA results (Table 3) indicate that there was ap-
Table 2. Results of Forward Stepwise Regression Data-Adjustment Procedures for Baseline Measures and 7-year Changes (ꢄ)
Mean Regression
Terms
Variance Regression
Terms
Parameter
%
%
3
Grip strength
Fathers
Mothers
Sons
Daughters
Fathers
Mothers
Sons
Daughters
Fathers
Mothers
Sons
Daughters
Fathers
Mothers
Sons
Daughters
Fathers
Mothers
Sons
Daughters
Fathers
Mothers
Sons
Daughters
Fathers
Mothers
Sons
Daughters
Fathers
Mothers
Sons
age , BMI
18.7
10.1
82.0
76.3
21.4
15.1
52.1
62.4
10.3
1.7
none
none
age, age
BMI
none
grip strength
none
none
none
none
none
none
none
flexibility
none
ꢄBMI, flexibility
age
none
—
—
9.9
3.9
—
3.1
—
—
—
—
—
—
—
1.5
—
—
0.8
—
3.4
—
4.4
3.0
—
—
0.9
—
—
—
—
—
—
—
3
age , BMI
2
2
age, age , BMI
2
3
age, age , age , BMI
2
ꢄ Grip strength
Trunk flexibility
ꢄ Trunk flexibility
Push-ups
age , ꢄBMI, grip strength
age, ꢄBMI, grip strength
age, ꢄBMI, grip strength
age, age , ꢄBMI
2
age
age
3
none
none
—
—
2
age , ꢄBMI, flexibility
17.7
15.3
11.3
4.6
26.5
8.5
3
age, age , ꢄBMI, flexibility
flexibility
flexibility
age, age , BMI
age, age
2
2
2
age, age
age
28.4
4.3
age
none
3
ꢄ Push-ups
Sit-ups
age, ꢄBMI, push-ups
15.2
44.0
32.3
38.4
43.4
37.9
22.3
—
age , push-ups
2
3
age , push-ups
age
push-ups
none
none
age
none
none
none
none
none
none
none
3
age , push-ups
age, BMI
age, BMI
2
3
age, age , age , BMI
none
3
ꢄ Sit-ups
age , ꢄBMI, sit-ups
5.9
age, ꢄBMI, sit-ups
16.6
40.1
34.0
age, sit-ups
3
Daughters
age , ꢄBMI, sit-ups
Note: BMI ꢁ body mass index.

B500
KATZMARZYK ET AL.
Table 3. Familial Aggregation of Baseline and 7-year Changes (ꢄ)
in Musculoskeletal Fitness Measures, Demonstrated by ANOVA
The significant familial resemblance in 7-year changes in
trunk flexibility and sit-ups could be explained partially by
shared environmental effects, as there is significant spousal
resemblance in addition to sibling or parent-offspring re-
semblance (Table 5). However, genetic factors may be more
important in explaining changes in grip strength and push-
ups, as the hypothesis of no spousal resemblance was not re-
jected, whereas hypotheses regarding no sibling and parent-
offspring resemblance were strongly rejected.
Table 6 presents the estimates of the familial correlations
under the most parsimonious models along with the herita-
bility estimates. Heritabilities were 64% for trunk flexibil-
ity, 37% for push-ups, 59% for sit-ups, and 48% for grip
strength. Similarly, heritabilities for the change scores were
Parameter
F
R2
P
Grip strength
ꢄ Grip strength
Trunk flexibility
ꢄ Trunk flexibility
Push-ups
ꢄ Push-ups
Sit-ups
ꢄ Sit-ups
1.41
1.41
1.83
1.68
1.64
1.18
2.08
1.43
0.48
0.58
0.55
0.63
0.53
0.54
0.59
0.59
ꢇ0.0001
ꢇ0.0001
ꢇ0.0001
ꢇ0.0001
ꢇ0.0001
0.05
ꢇ0.0001
0.0002
Notes: F is the ratio of between-family to within-family variance. ANOVA ꢁ
analysis of variance.
4
8% for trunk flexibility, 52% for push-ups, 41% for sit-
ups, and 32% for grip strength.
proximately 40–100% more variance between families than
within families, based on an examination of the F ratios.
The dependent variable, which in this case was family
membership (family identification number), accounted for
between 48% and 59% of the variance in adjusted baseline
measures and between 54% and 63% of the adjusted 7-year
changes. Thus, indicators of musculoskeletal fitness aggre-
gate significantly within the families of the CFS.
The results of the familial correlation model fitting proce-
dures are presented in Table 4 for the baseline measures and
in Table 5 for the 7-year changes. The hypothesis of no fa-
milial resemblance is strongly rejected for all phenotypes,
indicating that there is indeed familial resemblance in mus-
culoskeletal fitness both cross-sectionally and longitudi-
nally, confirming the ANOVA results. For grip strength,
trunk flexibility, and somewhat for push-ups, the pattern of
no spousal resemblance coupled with significant sibling and
parent-offspring resemblance suggests the role of genes in
explaining a portion of the familial resemblance. In con-
trast, the significant spousal resemblance in addition to sib-
ling and parent-offspring resemblance for sit-ups suggests
that shared environmental factors are important in explain-
ing this familial resemblance.
DISCUSSION
The present investigation distinguishes itself from previ-
ous studies of the familial resemblance in musculoskeletal
fitness by its longitudinal design. Unfortunately, the appar-
ent lack of previous longitudinal genetic analyses of muscu-
loskeletal fitness prevents the comparison of our results
with those of others. However, the cross-sectional heritabil-
ity estimates obtained in the present study compare well
with transmissibility estimates (from a tau path analysis)
obtained by Pérusse and colleagues (13) for the original
CFS sample (n ꢁ 13,804). Estimates of the transmissibility
from parents to offspring through both biological and cul-
tural paths for trunk flexibility, push-ups, sit-ups, and grip
strength were 48%, 44%, 37%, and 37%, respectively (13).
Thus, we have provided consistent evidence that muscu-
loskeletal fitness aggregates within the families of the origi-
nal CFS, a representative sample of the population.
Several studies have investigated familial resemblance in
muscular strength and endurance cross-sectionally by using
both family and twin-study designs. Twin studies generally
produce higher estimates of heritability than family studies.
An earlier twin study by Engström and Fischbein (16) dem-
Table 4. Summary of Results of Hypothesis Tests for Baseline Musculoskeletal Fitness Measures
Grip Strength
Trunk Flexibility
Push-ups
Sit-ups
Parameter
p*
AIC†
p*
AIC†
p*
AIC†
p*
AIC†
1
2
3
4
5
6
7
8
9
. General model
24.00
21.03
19.04
17.05
31.68
34.49
53.34
22.08
60.48
24.00
16.81
15.04
13.08
31.57
50.34
66.67
23.86
89.20
24.00
27.31
34.22
33.94
36.11
44.60
57.54
26.13
78.51
24.00
18.84
22.52
24.30
22.92
56.52
45.36
53.44
102.57
. No sex differences in offspring
. No sex differences in offspring or parents
. No sex or generation differences
. All correlations equal (environmental model)
. No sibling resemblance
. No parent-offspring resemblance
. No spouse resemblance
. No familial resemblance
0.28
0.41
0.54
0.94
0.96
0.98
0.02
0.001
0.001
0.0005
ꢇ0.0001
ꢇ0.0001
0.04
0.59
0.13
0.06
0.07
ꢇ0.0001
ꢇ0.0001
ꢇ0.0001
ꢇ0.0001
0.003
0.003
0.0009
ꢇ0.0001
0.78
ꢇ0.0001
ꢇ0.0001
0.17
ꢇ0.0001
ꢇ0.0001
ꢇ0.0001
10. Most parsimonious model
General model
ꢅ/ꢅ
24.00
Models 4 and 8
Model 2
0.64
15.13
0.88
13.06
0.59
18.84
2
*
p from the likelihood ratio ꢆ test; a significant value ( p ꢇ .05) indicates rejection of the hypothesis.
†
AIC ꢁ Akaike’s information criterion; the most parsimonious model is the one with the smallest AIC; we subtracted (ꢅ2 ln L) under the general model from the
AIC of all the submodels for easy comparison.

GENETICS OF MUSCULOSKELETAL FITNESS
B501
Table 5. Summary of Results of Hypothesis Tests for 7-year Changes (ꢄ) in Musculoskeletal Fitness Measures
ꢄ Grip Strength
ꢄ Trunk Flexibility
ꢄ Push-ups
ꢄ Sit-ups
Parameter
p*
AIC†
p*
AIC†
p*
AIC†
p*
AIC†
1
2
3
4
5
6
7
8
9
. General model
24.00
33.76
33.13
31.16
29.28
34.40
26.64
25.35
38.39
24.00
25.13
25.24
22.06
20.31
39.11
37.75
35.05
56.38
24.00
22.04
24.33
24.16
51.81
43.62
44.20
25.29
62.71
24.00
19.45
17.45
16.53
14.55
24.16
32.62
31.03
41.07
. No sex differences in offspring
. No sex differences in offspring or parents
. No sex or generation differences
. All correlations equal (environmental model)
. No sibling resemblance
. No parent-offspring resemblance
. No spouse resemblance
. No familial resemblance
0.001
0.002
0.004
0.007
0.0009
0.03
0.08
0.10
0.12
0.20
0.07
0.06
ꢇ0.0001
ꢇ0.0001
ꢇ0.0001
0.07
0.49
0.63
0.61
0.72
0.10
0.002
0.003
ꢇ0.0001
0.17
0.0001
0.0002
0.0003
ꢇ0.0001
0.07
0.0002
ꢇ0.0001
1
0. Most parsimonious model
General model
Model 5
ꢅ/ꢅ
24.00
0.17
20.31
0.72
14.55
Model 2
0.20
22.04
2
*
p from the likelihood ratio ꢆ test; a significant value (p ꢇ .05) indicates rejection of the hypothesis.
†
AIC ꢁ Akaike’s information criterion; the most parsimonious model is the one with the smallest AIC; we subtracted (ꢅ2 ln L) under the general model from the
AIC of all the submodels for easy comparison.
onstrated an F ratio of 4.28 between DZ and MZ twins for
an aggregate measure of muscular strength, adjusted for
body height. A more recent analysis of 10-year-old twins
estimated that 65% and 72% of the variance in trunk
strength and static arm-pull strength, respectively, are at-
tributable to genes (17). In contrast, estimates of genetic
heritability from the Québec Family Study (in which it was
possible to distinguish between the genetic and cultural
transmission) were 30% for muscular strength and 21% for
muscular endurance, whereas cultural inheritance accounted
for an additional 31% and 33% of the variance, respectively
there was both significant sibling resemblance and parent-
offspring resemblance in grip strength, coupled with no
spousal resemblance (Table 4), which suggests that genes
are responsible for explaining a portion of the familial re-
semblance in grip strength.
Flexibility is a joint-specific characteristic, and it is re-
lated to joint morphology. Indeed, the International Consen-
sus Document on Physical Activity, Fitness and Health in-
cludes flexibility in the “morphological” component of
health-related fitness (28). Thus, the relatively high esti-
mates of heritability for sit-and-reach flexibility obtained in
this study could partially be explained by the influence of
genes on the morphology (bones, tendons, and ligaments) of
the hip joint.
In summary, this study found significant familial resem-
blance for measures of musculoskeletal fitness, both cross-
sectionally and longitudinally, in the Canadian population.
The finding of significant heritability for changes in muscu-
loskeletal fitness over time suggests that there may be a ge-
netic susceptibility to the functional decline that is observed
with age. In contrast, approximately 40–70% of the vari-
(
14).
Estimates of transmissibility from parents to offspring for
both dominant grip strength and trunk flexibility in a sample
of Mennonite families were 0% and 66%, respectively (12).
There was, however, significant sibling resemblance in
dominant hand-grip strength that was almost completely ex-
plained by shared environmental effects. The zero trans-
missibility for dominant hand-grip strength is difficult to
explain, as it goes against uniformly high estimates of heri-
tability in other studies (8,14). Indeed, in the present study
Table 6. Estimates of Familial Correlations and Maximal Heritabilities Under the Most Parsimonious Model for Baseline and 7-year
Changes in Musculoskeletal Fitness Measures
Grip Strength
Trunk Flexibility
Push-ups
Sit-ups
Parameter
Baseline
7-y Change
Baseline
7-y Change
Baseline
7-y Change
Baseline
7-y Change
FM
FS
MS
FD
MD
SD
SS
DD
[0.00]
0.16 ꢈ 0.08
0.04 ꢈ 0.13
ꢅ0.06 ꢈ 0.11
0.39 ꢈ 0.12
0.22 ꢈ 0.11
ꢅ0.11 ꢈ 0.11
0.39 ꢈ 0.09
0.25 ꢈ 0.25
32%
[0.00]
0.27 ꢈ 0.04
[0.27 ꢈ 0.04]
[0.27 ꢈ 0.04]
[0.27 ꢈ 0.04]
[0.27 ꢈ 0.04]
[0.27 ꢈ 0.04]
[0.27 ꢈ 0.04]
[0.27 ꢈ 0.04]
48%
0.12 ꢈ 0.06
ꢅ0.03 ꢈ 0.11
0.22 ꢈ 0.07
0.18 ꢈ 0.09
0.45 ꢈ 0.06
0.34 ꢈ 0.07
0.29 ꢈ 0.12
0.17 ꢈ 0.13
37%
[0.00]
0.32 ꢈ 0.05
0.37 ꢈ 0.06
0.17 ꢈ 0.06
[0.37 ꢈ 0.06]
[0.17 ꢈ 0.06]
0.40 ꢈ 0.06
[0.40 ꢈ 0.06]
[0.40 ꢈ 0.06]
59%
0.22 ꢈ 0.04
0.24 ꢈ 0.03
[0.24 ꢈ 0.03]
[0.24 ꢈ 0.03]
[0.24 ꢈ 0.03]
[0.24 ꢈ 0.03]
[0.24 ꢈ 0.03]
[0.24 ꢈ 0.03]
48%
0.32 ꢈ 0.03
[0.32 ꢈ 0.03]
[0.32 ꢈ 0.03]
[0.32 ꢈ 0.03]
[0.32 ꢈ 0.03]
[0.32 ꢈ 0.03]
[0.32 ꢈ 0.03]
64%
0.07 ꢈ 0.31
0.38 ꢈ 0.10
ꢅ0.34 ꢈ 0.21
0.42 ꢈ 0.09
0.54 ꢈ 0.10
0.14 ꢈ 0.16
0.48 ꢈ 0.15
52%
[0.22 ꢈ 0.04]
[0.22 ꢈ 0.04]
[0.22 ꢈ 0.04]
[0.22 ꢈ 0.04]
[0.22 ꢈ 0.04]
[0.22 ꢈ 0.04]
[0.22 ꢈ 0.04]
41%
Max. Heritability
Notes: Estimates of familial correlations are ꢈSE; values in brackets are fixed or equal to a preceding value; the maximal heritability is computed as (rsib ꢂ rpꢅo
pꢅo). FM ꢁ father–mother; FS ꢁ father–son; MS ꢁ mother–son; FD ꢁ father–daughter; MD ꢁ mother–daughter; SD ꢁ son–daughter;
SS ꢁ son–son; DD ꢁ daughter–daughter.
)
(1 ꢂ rspouse)/(1 ꢂ rspouse ꢂ 2rspouser

B502
KATZMARZYK ET AL.
ance was unaccounted for, depending on the particular phe-
notype, which indicates that measures of musculoskeletal
fitness are not fixed but rather modifiable characteristics.
Recent North American physical activity recommendations
encourage strength-developing activities (resistance train-
ing) as a component of habitual physical activity (29,30).
Taken together, these results and recommendations suggest
that lifestyle factors such as physical activity are important
in maintaining fitness levels over time, but they must be
viewed against the background of genetic susceptibility.
Although we have shown strong evidence for familial re-
semblance in changes in musculoskeletal fitness in the Ca-
nadian population, these analyses should be replicated in
other populations to demonstrate the robustness of the results.
The finding of significant familial aggregation indicates the
need for molecular genetic studies aimed at identifying spe-
cific genes that are related to changes in musculoskeletal fit-
ness. Additionally, there is a need for more refined analyses
of household characteristics that may influence the ob-
served familial aggregation.
10. Malina RM, Mueller WH. Genetic and environmental influences on
the strength and motor performance of Philadelphia school children.
Hum Biol. 1981;53:163–179.
1
1. Szopa J. Familial studies on genetic determination of some manifesta-
tions of muscular strength in man. Genet Polonica. 1982;23:65–79.
12. Devor EJ, Crawford MH. Family resemblance for neuromuscular per-
formance in a Kansas Mennonite community. Am J Phys Anthropol.
1
984;64:289–296.
1
3. Pérusse L, Leblanc C, Bouchard C. Inter-generation transmission of
physical fitness in the Canadian population. Can J Sport Sci. 1988;13:
8–14.
14. Pérusse L, Lortie G, Leblanc C, Tremblay A, Thériault G, Bouchard
C. Genetic and environmental sources of variation in physical fitness.
Ann Hum Biol. 1987;14:425–434.
1
5. Komi PV, Klissouras V, Karvinen E. Genetic variation in neuromus-
cular performance. Int Z Angew Phys. 1973;31:289–304.
16. Engström L-M, Fischbein S. Physical capacity in twins. Acta Genet
Med Gemellol. 1977;26:159–165.
1
7. Maes H, Beunen G, Vlietinck R, et al. Heritability of health- and
performance-related fitness. In: Duquet W, Day JAP, eds. Kinanthro-
pometry IV. London: Spon; 1993:140–149.
18. Chatterjee S, Das N. Physical and motor fitness and twins. Jpn J Phys-
iol. 1995;45:519–534.
1
2
9. Thomis MA, Van Leemputte M, Maes HH, et al. Multivariate genetic
analysis of maximal isometric muscle force at different elbow angles.
J Appl Physiol. 1997;82:959–967.
0. Maes HH, Beunen GP, Vlietinck RF, et al. Inheritance of physical fit-
ness in 10-yr-old twins and their parents. Med Sci Sports Exerc. 1996;
28:1479–1491.
Acknowledgments
This research was supported by a grant from the Heart and Stroke Foun-
dation of Canada. Claude Bouchard is supported, in part, by the George A.
Bray Chair in Nutrition.
21. Fitness Canada. Fitness and Lifestyle in Canada. Ottawa, ON: Gov-
ernment of Canada; 1983.
Special thanks to Cora Craig and her colleagues at the Canadian Fitness
and Lifestyle Institute for making available the 1981 Canada Fitness Survey
and 1988 Campbell’s Survey databases.
2
2. Fitness Canada. Canadian Standardized Test of Fitness. Ottawa, ON:
Government of Canada; 1986.
2
3. SAS Institute. SAS/STAT Users Guide. Cary, NC: SAS Institute; 1988.
Address correspondence to Dr. Peter T. Katzmarzyk, School of Kinesi-
ology and Health Science, York University, 4700 Keele Street, North
York, Ontario M3J 1P3, Canada. E-mail: katzmarz@yorku.ca
24. Pérusse L, Rice T, Després J-P, Rao DC, Bouchard C. Cross-trait fa-
milial resemblance for body fat and blood lipids: familial correlations
in the Québec Family Study. Arterioscler Thromb Vasc Biol. 1997;17:
3
270–3277.
2
5. Province MA, Rao DC. General purpose model and computer program
for combined segregation and path analysis (SEGPATH): automati-
cally creating computer programs from symbolic language model
specifications. Genet Epidemiol. 1995;12:203–219.
6. Rao CR. Linear Statistical Inference and Its Applications. New York:
Wiley; 1973.
7. Akaike H. A new look at the statistical model identification. IEEE
Trans Automat Contr. 1974;19:716–723.
8. Bouchard C, Shephard RJ. Physical activity, fitness, and health: the
model and key concepts. In: Bouchard C, Shephard RJ, Stephens T,
eds. Physical Activity, Fitness and Health. Champaign, IL: Human Ki-
netics; 1994:77–88.
9. Canada’s Physical Activity Guide to Healthy Active Living. Ottawa,
ON: Health Canada; 1998.
0. Physical Activity and Health: A Report of the Surgeon General. At-
lanta, GA: US Department of Health and Human Services, Centers for
Disease Control and Prevention, National Center for Chronic Disease
Prevention and Health Promotion; 1996.
References
1
. Warburton DE, Gledhill N, Quinney A. Musculoskeletal fitness and
health. Can J Appl Physiol. 2001;26:217–237.
2
. Payne N, Gledhill N, Katzmarzyk PT, Jamnik V, Ferguson S. Health
implications of musculoskeletal fitness. Can J Appl Physiol. 2000;25:
2
2
2
1
14–126.
3
4
5
6
. Cunningham DA, Paterson DH, Himann JE, Rechnitzer PA. Determi-
nants of independence in the elderly. Can J Appl Physiol. 1993;18:
2
43–254.
. Skelton DA, Greig CA, Davies JM, Young A. Strength, power and re-
lated functional ability of healthy people aged 65–89 years. Age
Ageing. 1994;23:371–377.
. Kallman DA, Plato CC, Tobin JD. The role of muscle loss in the age-
related decline of grip strength: cross-sectional and longitudinal per-
spectives. J Gerontol Med Sci. 1990;45:M82–M88.
. Backman E, Johansson V, Hager B, Sjoblom P, Henriksson KG. Isomet-
ric muscle strength and muscular endurance in normal persons aged
between 17 and 70 years. Scand J Rehabil Med. 1995;27:109–117.
. Bouchard C, Malina RM. Genetics for the sport scientist: selected meth-
odological considerations. Exerc Sport Sci Rev. 1983;11:275–305.
. Bouchard C, Malina RM, Pérusse L. Genetics of Fitness and Physical
Performance. Champaign, IL: Human Kinetics; 1997.
2
3
7
8
9
. Kovar R. Sledovani podobnosti mezi rodici a jejich potomky v nek-
terych motorickych projech. Teorie a Praxe Telesne Vychovy. 1981;
Received December 21, 2000
Accepted July 17, 2001
2
9:93–98.
Decision Editor: John A. Faulkner, PhD