University of Southern California Dissertations and Theses

Longitudinal changes in physical activity and physical fitness: associations with blood pressure

(USC Thesis Other)

Longitudinal changes in physical activity and physical fitness: associations with blood pressure

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content LONGITUDINAL CHANGES IN PHYSICAL ACTIVITY
AND PHYSICAL FITNESS:
ASSOCIATIONS WITH BLOOD PRESSURE
by
Ruolan Liu
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(APPLIED BIOMETRY AND EPIDEMIOLOGY)
August 2003
Copyright 2003 Ruolan Liu
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 1417929
INFORMATION TO USERS
The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
®
UMI
UMI Microform 1417929
Copyright 2004 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089-1695
This thesis, written by
under the direction of h€. thesis committee, and
approved by all its members, has been presented to and
accepted by the Director of Graduate and Professional
Programs, in partialfulfillment of the requirements for the
degree of
* Director
Date AiigiiBt 1 2 . 2003
Thesis Committee
U JQ M qL / (M c & cJC *
Chair
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DEDICATION
To my little baby boy Shaun Liu, my husband Wei Liu, my parents, my parents
law, and my brother.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGMENTS
I would like to thank my Master’s Thesis Committee, Dr. Wendy Mack, Dr.
Stanley Azen, and Dr. Howard Hodis, for their time and support through my thesis
work. I am grateful to Dr. Wendy Mack, my committee chairman, for her patience,
guidance and assistance to accomplish my Master’s thesis. Special thanks goes to
Dr. Stanley Azen, for his valuable advice and encouragement throughout my
studying in biostatistics. Also I deeply appreciate the continuous support and
understanding from Dr. Michel Baudry, who is my mentor and advisor through my
research career in Neuroscience Ph. D program.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
DEDICATION..................................................................................................................ii
ACKNOWLEDGMENTS...............................................................................................iii
LIST OF TABLES.......................................................................................................... vi
LIST OF FIGURES....................................................................................................... viii
ABSTRACT.....................................................................................................................ix
INTRODUCTION.............................................................................................................1
METHODS........................................................................................................................ 5
Study design and subjects.....................................................................................5
Physical activity.................................................................................................... 7
Physical fitness...................................................................................................... 7
Blood pressure and other risk factors for cardiovascular diseases....................8
Statistical analyses................................................................................................9
RESULTS........................................................................................................................ 12
Baseline physical activity and physical fitness and their relationship with
blood pressure (cross-sectional analyses)..........................................................12
Longitudinal changes in physical activity and physical fitness and their
relationship with blood pressure (longitudinal analyses)................................ 17
1. Longitudinal analyses of physical activity and the association
with blood pressure........................................................................22
2. Longitudinal analyses of physical fitness and the association
with blood pressure......................................................................... 30
3. Evaluation of possible modifying effects of physical fitness
level and follow-up examination period on the association
between physical activity and blood pressure..............................33
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DISCUSSION.
REFERENCES
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES
Table 1. Baseline characteristics for the Los Angeles County Firefighter
Cohort......................................................................................................13
Table 2. Comparisons of SBP and DBP among physical activity and
physical fitness levels: Cross-sectional analysis from baseline
examination............................................................................................. 14
Table 3. Univariate linear regression for BP as dependent variables with
physical activity, physical fitness and other related predictors:
Cross-sectional analysis from baseline examination........................... 16
Table 4. Multiple 1 linear regression for BP as dependent variables with
physical activity, physical fitness and other related predictors:
Cross-sectional analysis from baseline examination........................... 17
Table 5. Comparisons of SBP and DBP among physical activity and
physical fitness levels at the end of follow-up (last examination):
Cross-sectional analysis.........................................................................20
Table 6. Longitudinal analysis of the relationship between physical activity
and other CVD risk factors: Dependent variables = physical fitness,
and BMI; independent variables = physical activity, time since
first examination, and activity * time interaction................................21
Table 7. Longitudinal analysis of the influence of smoking on physical
activity and blood pressure: Dependent variables = physical
activity, SBP and DBP; independent variables = smoking,
years since first examination, and interaction of smoking * years
since first examination........................................................................... 23
Table 8. Longitudinal analysis of the influence of age on physical activity
and blood pressure: Dependent variables = physical activity, SBP
and DBP; independent variables = age, years since first examination,
and interaction of age * years since first examination.........................25
Table 9. Longitudinal analysis of the relationship between levels of physical
activity and blood pressure: Dependent variables = SBP and DBP;
independent variables = physical activity, years since first
examination, and interaction term of activity * years since first
examination.............................................................................................26
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 10. Longitudinal analysis of the relationship between levels of
physical activity and blood pressures with adjustment for other
CVD risk factors: Dependent variables = SBP and DBP;
independent variables - physical activity, years since first
examination, and interaction term of activity * years since first
examination.............................................................................................28
Table 11.
Table 12.
Longitudinal analysis of the relationship between the levels of
physical fitness and blood pressure: Dependent variables = SBP
and DBP; independent variables = physical fitness level,
years since first exam, and the interaction of the tw o .................. .31
Longitudinal analysis on the relationship between the levels of
physical fitness and blood pressure with adjustment for other CVD
risk factors: Dependent variables = SBP and DBP; independent
variables = physical fitness level, years since first exam, and the
interaction of the two............................................................................. 32
Table 13. Longitudinal analysis of the univariate relationship between
physical activity and SBP by physical fitness level: Dependent
variable = SBP; independent variables - physical activity, years
since first exam, and the interaction of the two.............................. .36
Table 14. Longitudinal analysis of the univariate relationship between
physical activity and DBP by physical fitness level: Dependent
variable = SBP; independent variables = physical activity, years
since first exam, and the interaction of the two.............................. .37
Table 15. Longitudinal analysis of the relationship between physical activity
and SBP by follow-up measurement periods: Dependent variable
= SBP; independent variables = physical activity, years since first
exam, physical fitness, and the interaction of the physical
activity/fitness with years since first exam...........................................39
Table 16. Longitudinal analysis of the relationship between physical activity
and DBP by follow-up measurement periods: Dependent variable
= DBP; independent variables = physical activity, years since first
exam, physical fitness, and the interaction of the physical
activity/fitness with years since first exam...........................................40
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF FIGURES
Figure 1. Comparison of physical fitness between first and last
measurement...........................................................................................34
Figure 2. Longitudinal change in physical fitness between first and last
measurement...........................................................................................34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
The objective of this study was to investigate the longitudinal relationships
among physical activity, physical fitness and blood pressure, and to examine the
intensity and duration effects of physical activity on blood pressure. Subjects were
2940 males from Los Angels County firefighter cohort, an observational longitudinal
study started in 1971 with a maximum follow-up period of 20 years. Physical
activity, physical fitness, blood pressure, age, BMI and smoking behavior were
repeatedly measured during follow-up. The longitudinal analyses indicated that
increasing physical activity and fitness were both inversely related to rates of change
in blood pressure. The relationships were consistently observed in different follow-
up periods.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INTRODU CION
Cardiovascular diseases (CVD) are the most common cause of death for both
men and women among all racial and ethnic groups, and account for more than 40%
of all deaths in the United States (data from American Heart Association, 2002 Heart
and Stroke Statistical Update). According to the American Heart Association, at
least 58 million Americans have some form of CVD, including coronary heart
disease, high blood pressure, and others. Extensive clinical and observational studies
have identified several risk factors for CVD. Despite some important risk factors,
such as age, gender and race, which cannot be modified, three modifiable health-
related behaviors (smoking, insufficient physical activity, and diet) contribute greatly
to the incidence of cardiovascular diseases (data from Centers for Disease Control
and Prevention, 2000 Behavioral Risk Factor Surveillance System). Since the 1950s
increasing evidence from observational studies and clinical trials indicates that a lack
of physical activity, either in occupational or leisure time, is a major independent risk
factor for CVD. It has been estimated that the population attributable risk (PAR) of
physical inactivity for CVD mortality is 35% (Powell and Blair 1994), which means
that 35% of deaths caused by CVD could be prevented if everyone had a sufficiently
physically active lifestyle. Unfortunately, 26.9% of U.S. adults are sedentary, a
prevalance similar to smoking (23.2%) (data from Centers for Disease Control and
Prevention, 2000 Behavioral Risk Factor Surveillance System).
Physical activity is crucial to physical and mental health, as well as social
well-being. Physical activity can be defined as voluntary skeletal muscle movements
1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
to expend energy, and one of its major benefits is to improve levels of physical
fitness. Physical fitness is usually measured as maximal aerobic capacity to
integratively measure both cardiorespiratory and neuromusculo-skeletal function,
and is mainly determined by physical activity patterns over recent weeks or months
(Erikssen 2001; Williams 2001). A large body of epidemiological studies
consistently suggests that leisure time physical activity or physical fitness is
inversely correlated with the risk of cardiovascular diseases (Powell et al. 1987;
Katzmarzyk et al. 1999; Lefevre et al. 1999; Twisk et al. 2002). Poor
cardiorespiratory fitness has been also shown to significantly increase the risk of
cardiovascular mortality (Kavanagh 2001).
Generally, increases in physical activity result in increases in physical fitness,
although the amount of change in fitness relative to a given change in exercise levels
varies widely from person to person (Erikssen 2001; Williams 2001). Although
physical activity and physical fitness are closely related, they are usually thought as
separate risk factors for cardiovascular diseases (Blair et al. 2001; Blair and Jackson
2001). All measures of physical fitness (i.e., cardiorespiratory endurance, muscular
strength, flexibility, and body composition) have genetic components that are also
strongly influenced by environmental factors, and the specific fitness levels achieved
by an individual are determined by genetic makeup, environmental factors, and
genetic-environmental interactions (Blair and Jackson 2001).
Physical fitness is sometimes treated as a surrogate measurement of health-
related behaviors such as physical activity, and this health-related fitness is one
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
important determinant for various health-related outcomes. Some studies suggest
that physical activity must reach a certain threshold (specifically, regular and
vigorous aerobic exercise) to be effective in reducing CVD risk, and to be associated
with physical fitness level (Lee et al. 1995; Blair et al. 2001). However, few data are
available evaluating the joint effects of physical activity and physical fitness.
Whether physical activity must be present with some threshold levels of physical
fitness to achieve cardiovascular benefits has not been clearly answered. In addition,
it is not known whether beneficial effects of physical activity and/or physical fitness
are short-term or long lasting. Therefore, the relationship between physical activity
and physical fitness and cardiovascular outcomes still needs to be further
investigated.
Physical activity and physical fitness modify a variety of CVD risk factors
including blood pressure, resting heart rate, total cholesterol, body mass index,
prevalence of smoking, and others (Powell et al. 1987; Church et al. 2001; Kavanagh
2001). Blood pressure (especially at levels definitive for hypertension) is an
influential etiologic factor for the incidence of important cardiovascular events, such
as myocardial infarction, stroke, and sudden coronary death. The association of
lifestyle factors with high blood pressure has been long-recognized. Regular
exercise has the potential to reduce blood pressure, especially in hypertensive
patients and in subjects with high-normal blood pressure (Lesniak and Dubbert 2001;
Rankinen and Bouchard 2002). Due to inconsistency in the design, analysis and
reporting of many studies, the meta-analyses of such studies relating physical
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
activity to blood pressure reduction has produced mixed results. The ranges and
levels of both systolic and diastolic blood pressure reductions varied from study to
study. Some studies showed that the age-adjusted risk for hypertension was
significantly higher in sedentary, least active men, whereas others failed to find a
significant association between blood pressure and physical activity (or physical
fitness) (Lesniak and Dubbert 2001). It is commonly thought that mild to moderate
intensity exercise are more effective in lowering blood pressure than strenuous
intensity exercise (Lesniak and Dubbert 2001). However, the optimal levels of
exercise intensity, frequency, and duration for reaching optimal levels of physical
fitness to achieve cardiovascular benefits such as blood pressure reduction require
further identification. The full relationship among blood pressure, physical activity
and physical fitness has not been completely evaluated.
The cardiovascular morbidity and mortality experiences of firefighters has
long been of research and practical interest because firefighters are potentially
exposed to acute psychological, physical, and environmental stresses (i.e., alarms,
smoke, and a vast array of toxic substances) in their job duties. While it has been
hypothesized that firefighters would have higher incidence and mortality rates for
CVD, many well-conducted cohort and case-control studies do not support this
assumption and instead suggest a strong healthy worker effect (Guidotti 1992;
Horowitz and Montgomery 1993; Tornling et al. 1994). The evidence from
Canadian firefighter and Los Angeles County firefighter cohort studies indicate that
firefighters are not distinctly different from the general population, and have no
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
significantly elevated or lowered cardiovascular risk factors compared to the general
population (Horowitz and Montgomery 1993). On the other hand, these firefighters
must perform intense exercise in arduous environments, and job requirements often
routinely assess their physical work capacity and other cardiovascular risk factors.
Therefore, this subgroup of the general population is an interesting source to study
the relationships between physical activity and physical fitness and other
cardiovascular risk factors.
The large longitudinal database used in the present study came from a Los
Angeles County firefighter cohort, and will be analyzed to address the following
objectives:
1. To investigate the cross-sectional and longitudinal relationships of physical
activity and physical fitness associated with systolic and diastolic blood
pressure.
2. To test the modifying effects of physical fitness on the relations between
physical activity and systolic and diastolic blood pressure.
3. To examine the duration of physical activity with respect to the relationship
between physical activity and systolic and diastolic blood pressure.
METHODS
Study design and subjects
The data were derived from the Los Angeles County firefighter cohort study,
an observational longitudinal study that began in 1971 as part of routine assessment
5
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of physical fitness and cardiovascular risk factors by the Los Angeles County
Department of Human Resources. Because of the strenuous physical requirements
of firefighting, these factors were assessed at routine intervals (approximately every
three years) in both prospective and current firefighters of the Los Angeles County
Fire Department. This cohort included all 3309 firefighters evaluated from 1971 to
1991. The exercise capacity measurements were conducted at the Cardiopulmonary
Laboratory of the Occupational Health Service, Department of Human Resources,
Los Angeles County. In addition to the variables of primary interest, anthropometric
parameters (e.g., body weight, body height) and biologic parameters (e.g., age, sex,
lipoprotein levels, blood pressure) were measured.
Preliminary results from this cohort indicated that increased cardiovascular
disease risk was significantly associated with low physical work capacity, especially
for subjects who had higher levels of other cardiovascular risk factors, such as high
blood pressure, smoking, and high cholesterol. For the present study, women (n=l 1)
were excluded because over 99% of the firefighters were men. In addition, men with
either height lower than 155 cm or weight greater than 120 kg (n=358) were also
excluded because they might have unique physical characteristics influencing the
relationship between physical fitness and activity and blood pressure. The remaining
2940 male firefighters were used to examine the longitudinal relationships between
physical activity and physical work capacity and blood pressure.
Over a follow-up period of 20 years, subjects were longitudinally evaluated
to a maximum of 11 repeated measurements. At least four repeated measurements
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
were carried out in more than 50% of the subjects and the mean follow-up period
was 8.3 years. The initial age of the subjects at the beginning of the longitudinal
study was 33.1 (±8.2) years, and the average age at the end of follow-up was 44.4 (±
9.4) years.
Physical activity
Physical activity is a muscular body movement to cause energy expenditure,
and this energy expenditure varies widely between individuals to affect one’s daily
metabolic rate. Daily physical activity is usually measured by a structured
questionnaire or interview, and the physical activity level is measured as the number
of days per week that the individual reports a continuous strenuous exercise lasting at
least 30 minutes (Blair et al. 2001; Kohl 2001). In the present study, a six-level scale
of physical activity was used as a simple and quick estimate of activity, ranging from
sedentary, 1 day per week, up to 5 days per week. Physical activity measured by this
method showed a moderate correlation with the physical work capacity measure used
in this study (Pearson’s correlation coefficient p = 0.35, p < 0.01), and this
correlation was consistent with previously reported activity measures (Seefeldt et al.
2002).
Physical fitness
In this dataset, the assessment of physical fitness included indicators of
cardiorespiratory fitness (physical work capacity), muscular fitness (grip, leg, and
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
back muscle strength measures), and morphological fitness (four skinfold
measurements at biceps, triceps, suprailiac, and subscapular sites). Only
cardiorespiratory fitness was selected for analysis in the present study.
Physical work capacity was measured by bicycle ergometry during a 20-
minute heart-rate-controlled graded exercise test. The average power output (in
watts) at the target heart rate of 160 beats per minute was used as the primary
physical fitness measure. This measure of physical fitness, like the similar measures
based on treadmill or bicycle ergometry used in other cohort studies, is a function of
aerobic fitness, muscle strength, and muscle endurance. Physical fitness was used as
a continuous variable in most of the following analyses. To evaluate the effects of
different levels of physical fitness on the relations among physical activity and blood
pressure and other cardiovascular risk factors, physical fitness was further
categorized into 4 levels from low to high magnitude according to its quartile
distribution.
Blood pressure and other risk factors for cardiovascular diseases
Blood pressure is an important dimension of cardiovascular health status.
Systolic (SBP) and diastolic (DBP) blood pressures were measured with a standard
pressure cuff placed around the upper arm. With a sphygmomanometer, SBP and
DBP were recorded as the subject at rest in the sitting position. If subjects were
taking antihypertensive medications, they were required to cease their medication for
two days before the examination. Subjects were categorized as hypertensive if they
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
had a blood pressure (SBP/DBP) of 140/90 mm Hg or higher, as high normal blood
pressure if their blood pressure was greater than 120/80 mm Hg but less than 140/90
mm Hg, and normal if their blood pressure was 120/80 mm Hg or less.
Obesity is associated with cardiovascular and other metabolic diseases.
Overweight and obesity are clinically defined by body mass index (BMI). The BMI
(kg/m2 ) was calculated by dividing body weight measured in kilograms by the height
in meters squared. In this study, body weight was measured in underwear, and
height was measured without shoes. Subjects were categorized as obese if their BMI
9 9
was greater than 30 kg/m , as overweight if their BMI was between 25-29.9 kg/m ,
and normal weight if their BMI was less than 25 kg/m2.
Smoking behavior is another important health indicator. At each of the
examinations, smoking status was measured as one of three groups (never, quit, and
current). The levels of total cholesterol and total triglycerides, glucose, and other
cardiovascular risk factors were also measured in this cohort, but will not be used in
the present study.
Statistical analyses
These data exist on a SAS (Statistical Analysis System software) database.
For all statistical analyses, a significance level of p < 0.05 was used and all analyses
were carried out using the SAS statistical software package.
The relationships of physical activity and physical fitness with SBP and DBP
were assessed cross-sectionally using the measurements from each subject’s first
9
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
examination. The baseline (first examination) SBP and DBP were compared
among the levels of physical activity and the levels of physical fitness by using
analysis of variance (ANOYA) and Tukey’s multiple comparison test. Among
physical activity or physical fitness categories, the linear or quadratic trends of SBP
and DBP were tested using general linear model (GLM) procedures.
The cross-sectional associations of variables measured at the baseline
examination were also evaluated using physical fitness and activity as continuous
measures. First, physical activity or physical fitness was correlated to blood pressure
or other health parameters separately by univariate linear regression analysis.
Second, the associations between physical activity and physical fitness and blood
pressure were analyzed in a multivariate linear regression model including other
potential confounders (i.e., age, BMI, smoking behavior).
A paired t-test was used to test whether the difference in physical activity and
physical fitness between the first and last measure was significantly different, and
Pearson’s correlation analysis was performed to test if there was a correlation
between change in physical fitness and the levels of physical activity at first or last
measure.
Since multiple repeated measures were taken over time on each subject, the
longitudinal relationships among physical activity and physical fitness and blood
pressure (SBP or DBP) were assessed with random effects models. The advantage
of random effects modeling for longitudinal data is that all data are used in one
regression model, resulting in one regression coefficient to indicate the
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
interrelationships among these variables (Verbeke and Molenberghs 2000). The
mixed linear model is a generalization of a standard linear model, and it does not
assume that an equal number of repeated measures should be taken from each
individual or that all individuals should be measured at the same time points
(Verbeke and Molenberghs 2000). First, the relationship between longitudinal
change in physical activity or in physical fitness and blood pressure (SBP or DBP)
were univariately analyzed using the mixed linear model. The blood pressure
variables were the dependent measures, with physical activity or fitness modeled as
independent variables. Secondly, a multiple longitudinal analysis was carried out, in
which time since first examination measure (variable will be defined as ‘Years’ in
analysis), chronological age, BMI, and smoking status were considered as the time-
dependent covariates in the mixed linear model.
The possible modifying effects of physical fitness levels (1 to 4 levels from
low to high magnitude of physical fitness) on longitudinal change in physical activity
associated with blood pressure and the effects of different follow-up periods (starting
from the first examination, short term defined as years 1-2, medium term as years 3-
5, and long term as years 6-10 and years 11-20) on the association between
longitudinal change in physical activity and physical fitness and blood pressure were
also investigated using the mixed linear model.
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
RESULTS
Baseline physical activity and physical fitness and their relationship with blood
pressure (cross-sectional analyses)
The baseline values of all parameters for each subject were defined as the
values at their first examination. The baseline levels of systolic and diastolic blood
pressure, physical activity and physical fitness, and demographic variables are
described in Table 1. Among the total 2940 subjects, the mean age was 33 years old,
2.8% were obese, and 60.4% had never smoked. The average SBP and DBP were
123 mm Hg and 77 mm Hg respectively, and 4.9% of subjects were categorized as
hypertensive.
Baseline blood pressure was compared among levels of physical activity or
by the levels of physical fitness (Table 2). The mean baseline SBP was significantly
different among the levels of physical activity (p-value for ANOVA test < 0.01).
Tukey’s multiple comparison tests showed that persons who exercised 2 days per
week had a significantly lower SBP than those who had a sedentary lifestyle or who
exercised 3 or more days per week (p < 0.05). The mean SBP for persons who
exercised 1 or 2 days per week were lower than those who were physically inactive
or who exercised 4 or 5 days per week (p-value for quadratic trend < 0.01). The
mean baseline DBP did significantly differ among the different physical activity
levels (p-value for ANOVA test =0.03), but there was no clear pattern of differences
by physical activity level.
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 1. Baseline characteristics for the Los Angeles County firefighter cohort
Continuous variable N Percentage Mean ± SD Range
SBP (mmHg) 2940 123.1 ± 14.0 90-183
DBP (mmHg) 2940 76.7 ± 9.3 48-110
Physical fitness (watts)* 2940 152.3 ± 35.3 21 - 316
Age (years) 2940 33.1 ±8.1 17-59
BMI (Kg/m2 ) 2940 25.0 ±2.3 17.4-38.1
Categorical variable
Physical activity
Sedentary 105 3.6%
1 day / week 464 15.8%
2 days / week 821 27.9%
3 days / week 1041 35.4%
4 days / week 258 8.8%
5 days / week 251 8.5%
Smoking status
Never 1777 60.5%
Quit 365 12.4%
Current 798 27.1%
BP
Normal: < 120/80 mmHg 1338 45.5%
High normal: 120/80 ~ 140/90 mmHg 1459 49.6%
Hypertension: > 140/90 mmHg 143 4.9%
BMI
Normal: < 25 Kg/m2 1542 52.5%
Overweight: 25 ~ 29.9 Kg/m2 1315 44.7%
Obese: >30 Kg/m2 83 2.8%
• SBP = Systolic Blood Pressure, DBP - Diastolic Blood Pressure, BMI = Body Mass Index.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 2. Comparisons of SBP and DBP among physical activity and physical fitness levels: Cross-sectional analysis from baseline
examination
Subject SBP DBP
N
(% )
( Mean ± SD )
( Mean ± SD )
Physical activity 2940 100% 123.1 ± 14.0 76.7 ± 9.3
Level:
Sedentary 105 3.6% 126.5 ± 14.2* a 78.4 ±9.9
1 day / week 464 15.2% 122.3 ± 14.3 * b 76.3 ±9.1
2 days / week 821 27.9% 121.1 ± 13.2* a 76.7 ± 9.3
3 days / week 1041 35.4% 123.2 ± 14.0* a 76.4 ± 9.0
4 days / week 258 8.8% 126.7 ± 13.8* ab 78.2 ± 9.9
5 days / week 251 8.5% 125.9 ± 14.7* ab 76.5 ± 9.7
ANOVA test P<.0001 P = 0.03
Linear trend P = 0.17 P = 0.49
Quadratic trend P < .0001 P = 0.20
Physical fitness 2940 100% 123.1 ± 14.0 76.7 ±9.3
Level:
< 135 watts 874 29.7% 123.6 ± 13.9 77.4 ± 9.7 * a
135 ~ 160 watts 893 30.4% 122.0 ± 14.0 * a 75.9 ± 9.0 * a
160 ~ 185 watts 670 22.8% 123.0 ± 13.9 76.6 ± 9.3
>185 watts 503 17.1% 124.3 ± 14.2 * a 77.0 ±9.1
ANOVA test P = 0.01 P = 0.01
Linear trend P = 0.21 P = 0.72
Quadratic trend P = 0.006 P = 0.01
• ANOVA and Tukey’s multiple comparison tests were used.
• * BP groups within physical activity or within physical fitness which share a superscripted letter were significantly different
at p < 0.05.
• General linear model (GLM) procedure was used for testing linear and quadratic trends.
£
Among the quartile categories of physical fitness, the mean SBP in the
second quartile group was significantly lower than in the highest quartile of physical
fitness (p-value for Tukey’s test < 0.05), and the mean DBP in this second quartile
fitness group was also significantly lower than the mean DBP in the lowest quartile
of fitness (p-value for Tukey’s test < 0.05). Quadratic trends in the BP means were
significant among the levels of physical fitness for both SBP and DBP (p = 0.006; p
= 0.010, respectively).
The results of baseline univariate linear regression analyses relating physical
activity, physical fitness, or other important risk factors to blood pressure are
summarized in Table 3. Physical activity levels were positively associated with
physical fitness (regression coefficient ( 3 = 8.90, p < 0.01, data not shown). In a
model with a linear trend term only, physical activity was positively linearly related
with SBP (p = 0.85, p < 0.01), whereas there was no significant linear relationship
between physical fitness and SBP. Moreover, no significant linear associations
between physical activity and physical fitness and DBP were observed. In the
analyses including both linear and quadratic components for trend in blood pressure
by levels of physical fitness, the significant inverse linear trend (p<0.001 for both
SBP and DBP) indicated a reduction in blood pressure with increasing physical
fitness. The significant positive quadratic trend (p<0.001 for both SBP and DBP)
indicated that the mean blood pressure plateaued (and even increased, Table 2) with
high levels of physical fitness. This same pattern of association was observed for
physical activity and SBP, but not DBP (Table 3).
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 3. Univariate linear regression for BP as dependent variables with physical activity, physical fitness and
other related predictors: Cross-sectional analysis from baseline examination
SBP DBP
Parameter in Model
( 3
p-value r2 adjxr2
P
p-value r2
"V H
X
-a1
CS
Physical activity 0.85 < .0001 0.01 0.01 0.03 0.84 0.000 0.000
Physical activity -3.39 0.001 0.02 0.01 -0.71 0.29 0.0005 -0.0002
Physical activity2 0.58 <.0001 0.10 0.26
Physical fitness 0.005 0.53 0.000 0.000 -0.01 0.10 0.001 0.001
Physical fitness -0.16 < .0001 0.006 0.005 -0.14 <.0001 0.009 0.009
Physical fitness2 0.0005 < .0001 0.0004 < .0001
Age -0.19 < .0001 0.01 0.01 0.17 < .0001 0.02 0.02
BMI 0.56 <.0001 0.01 0.01 0.85 <.0001 0.05 0.05
Smoking status3 -2.33 < .0001 0.02 0.02 -0.63 0.001 0.004 0.003
• Linear regression analysis was used.
• a Current smoking status is reference category.
o
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 4. Multiple linear regression for BP as dependent variables with physical activity, physical fitness and other related
predictors: Cross-sectional analysis from baseline examination
Model Variables in model SBP DBP
P
p-value r2 adjxr2
(3
p-value r2 adjxr2
#1
Physical activity -3.06 0.002 0.02 0.01 0.06 0.66 0.009 0.008
Physical activity2 0.54 <.0001
Physical fitness -0.15 0.0004 -0.14 < .0001
Physical fitness2 0.0005 0.0006 0.0004 < .0001
#2
Age -0.26 <.0001 0.03 0.03 0.10 < .0001 0.05 0.05
BMI 0.83 < .0001 0.74 < .0001
#3
Physical activity -2.45 0.01 0.04 0.04 0.44 0.003 0.07 0.06
Physical activity2 0.45 0.001
Physical fitness -0.15 0.0002 -0.15 < .0001
Physical fitness2 0.0004 0.001 0.004 < .0001
Age -0.23 < .0001 0.11 < .0001
BMI 0.88 < .0001 0.79 < .0001
#4
Physical activity -2.67 0.007 0.06 0.05 0.37 0.01 0.07 0.07
Physical activity2 0.46 0.001
Physical fitness -0.14 0.001 -0.14 < .0001
Physical fitness2 0.0004 0.003 0.0004 < .0001
Age -0.19 <.0001 0.13 < .0001 < .0001
BMI 0.88 < .0001 0.78 < .0001
Smoking status3 -1.90 < .0001 -0.99 < .0001
• a Current smoking status is reference category
Age was significantly positively linearly associated with DBP (P = 0.17; p <
0.01), and was inversely linearly related to SBP (P = -0.19, p < 0.01). BMI were
significantly positively linearly related to both DBP and SBP (P = 0.85, p < 0.01; P =
0.56, p < 0.01, respectively). Smoking was inversely linearly related to both DBP
and SBP (P = -0.63, p < 0.01; P = -2.33, p < 0.01, respectively).
Several multiple linear regression models were evaluated (Table 4). The
quadratic relationship of initially decreasing blood pressure with moderate levels of
activity/fitness and increasing blood pressure at higher levels of activity/fitness were
still apparent for physical fitness (both for SBP and DBP) and physical activity (SBP
only). Physical activity was significantly linearly positively associated with DBP
after adjustment for age, BMI, and smoking. Physical fitness and physical activity
were significantly independently related to SBP and DBP. In the full multivariate
model, age, BMI, smoking, physical fitness, and physical activity were all
independently and significantly related to SBP and DBP.
Longitudinal changes in physical activity and physical fitness and their
relationship with blood pressure (longitudinal analyses)
In this Los Angeles County firefighter cohort, 10987 observations were
obtained over a maximum 20-year follow-up period. Fifty percent of the total 2940
subjects had at least 4 repeated measurements, and the mean follow-up period was
8.3 years.
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
At the end of the follow-up (each subject’s last examination), the average
SBP and DBP were 129 mm Hg and 82 mm Hg (Table 5). Inverse linear trends in
the mean SBP and DBP were statistically significant with increasing level of
physical activity (SBP: p = 0.008; DBP: p < 0.001) and also with increasing level of
physical fitness (SBP: p = 0.002; DBP: p < 0.001). The quadratic trends of
increasing mean SBP noted with higher levels of physical activity and physical
fitness measured in the baseline examination were not apparent at the last
examination.
The longitudinal relationships between physical activity and physical fitness
are displayed in Table 6 (first column of results). The mixed random effects model
was used with physical fitness as the dependent variable. Physical activity, years
since first examination, and an interaction term of activity * years since first
examination were included as independent variables. The main effect of physical
activity tests the cross-sectional relationship between level of physical activity and
fitness, the main effect term for years estimates the annual rate of change in fitness in
the referent activity group of 5 days per week, and the activity * years interaction
term evaluates whether rates of change in fitness vary by level of physical activity.
Cross-sectionally, lower levels of physical activity were associated with reduced
levels of physical fitness (p<0.001). The average rate of change in fitness in subjects
who were physically active 5 days per week was an increase of 2.25 watts per year (p
< 0.001). The rate of change in fitness significantly varied by level of physical
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5. Comparisons of SBP and DBP among physical activity and physical fitness levels at the end of follow-up (last
examination): Cross-sectional analysis
Subject SBP DBP
N (%) ( Mean ± SD) ( Mean ± SD )
Physical activity 2940 100.0 129.2 ± 13.7 82.2 ± 8.9
Level:
Sedentary 285 9.7 129.6 ± 13.1 83.9 ± 8.8 * a d e
1 day / week 466 15.8 130.5 ± 15.0 83.1 ± 9.2 * c g
2 days / week 660 22.4 130.0 ± 13.9 83.2 ±9.3 * bf
3 days / week 1082 36.8 128.6 ± 13.5 81.3 ± 8.6 * abc
4 days / week 290 9.9 128.4 ± 12.6 80.9 ± 8.2 * dfg
5 days / week 157 5.3 127.1 ± 12.7 81.0 ± 8.5 *e
ANOVA test P = 0.02 P< 0.001
Linear trend P = 0.008 P< 0.001
Quadratic trend P = 0.17 P = 0.58
Physical fitness 2940 100.0 129.2 ± 13.7 82.2 ± 8.9
Level:
< 135 watts 565 19.2 130.3 ± 14.7* a 83.4 ± 9.4 * a
135 ~ 160 watts 680 23.1 129.9 ± 14.6 82.5 ± 8.9 * b
160- 185 watts 757 25.8 129.0 ± 13.2 82.1 ±9.0
>=185 watts 938 31.9 128.2 ± 12.8* a 81.4 ± 8.4 *ab
ANOVA test P = 0.02 P = 0.0003
Linear trend P = 0.002 P< 0.001
Quadratic trend P = 0.67 P = 0.90
• ANOVA and Tukey’s multiple comparison tests were used.
• *BP groups within physical activity or within physical fitness which share a superscripted letter were significantly
different at p < 0.05.
• General linear model (GLM) procedure was used for testing linear and quadratic trends.
K >
o
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 6. Longitudinal analysis of the relationship between physical activity and other CVD risk factors
(Dependent variables = physical fitness, and BMI; independent variables = physical activity, time since first
examination, and activity * time interaction)
Physical fitness BMI
Estimate p-value Estimate p-value
Intercept 165.86 <.001 24.46 <.001
Physical activity (levels)3 < .001b < ,001b
Sedentary -24.09 <.001 0.61 <.001
1 day / week -20.33 <.001 0.62 <.001
2 days / week -17.12 <.001 0.63 <.001
3 days / week -10.49 <.001 0.50 <.001
4 days / week -4.77 0.007 0.24 0.005
5 days / week 0 0
Years 2.25 < .00 lb 0.13 < ,001b
Years* Physical activity (levels)3 0.013b 0.10b
Sedentary -0.81 0.003 0.02 0.10
1 day / week -0.62 0.005 0.01 0.26
2 days / week -0.49 0.016 0.001 0.91
3 days / week -0.37 0.06 -0.004 0.71
4 days / week -0.20 0.39 0.003 0.80
5 days / week 0 0
• A mixed random effects linear model was used.
• a The highest level of physical activity (5 days / week) is the reference category.
• b p-value from Type 3 tests of fixed effects.
activity (p = 0.013), with less active subjects experiencing significantly lower rates
of change relative to the most active subjects.
1. Longitudinal analyses of physical activity and the association with blood
pressure
Age, BMI, and smoking are potential confounding factors that could
influence the relation between physical activity and blood pressures. Therefore, the
longitudinal relationships between physical activity and age, BMI, and smoking were
assessed and are displayed in Tables 6, 7 and 8. Cross-sectionally, the levels of
physical activity were inversely related to BMI (Table 6, p < 0.001) and the average
rate of change in BMI in subjects who were physically active 5 days per week was
an increase of 0.13 watts per year (Table 6,p<0 .001). However, the annual rate of
change in BMI did not vary by physical activity level (Table 6, p = 0.10).
Smoking was cross-sectionally inversely associated with physical activity
and systolic and diastolic blood pressures (Table 7,p<0 .001 for all three models).
Compared with current smokers, non-smokers were more likely to be more
physically active (p < 0.001). However, non-smokers’ mean systolic and diastolic
blood pressures were higher than current smokers (p < 0.001). The rate of change in
physical activity did not differ by smoking group (Table 7, p=0.22), but annual rates
of increase in systolic and diastolic blood pressures were significantly lower in non-
smokers relative to current smokers (Table 7, p<0.001 for both blood pressure
models).
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 7. Longitudinal analysis of the influence of smoking on physical activity and blood pressure: Dependent variable =
physical activity, SBP, DBP; independent variables = smoking, years since first examination, and interaction of
smoking * years since first examination.
Physical activity SBP DBP
Estimate p-value Estimate p-value Estimate p-value
Intercept 3.26 <. 001 120.06 <.001 75.70 <.001
Smoking status3 <.001b <.001b <.001b
Never 0.42 <. 001 4.11 <.001 1.47 <.001
Quit 0.44 < 001 3.32 <.001 2.21 <.001
Current 0 0 0
Years -0.02 0.001b 0.85 < .001b 0.70 < ,001b
Years * Smoking status3 0.22b <.001b 0.002b
Never 0.01 0.25 -0.33 <.001 -0.14 0.001
Quit -0.002 0.73 -0.21 0.005 -0.16 0.001
Current 0 0 0
A mixed random effects linear model was used.
a Current smoking status is reference category.
b p-value from Type 3 tests of fixed effects.
t o
U >
Cross-sectionally, age was inversely related to physical activity (Table 7,
p<0.001), but the annual rate of change in physical activity did not differ by age
groups (p=0.08). Cross-sectionally, subjects in age decades 30-40 and 40-50 had
significantly lower SBP and DBP than subjects aged 50 and older. Subjects younger
than age 30 did not differ from subjects aged 50 and older on SBP, but did have
significantly lower DBP. Longitudinally, subjects aged 30-40 and 40-50 had
significantly greater rates of increase in both SBP and DBP compared to subjects
aged 50 and older. Subjects younger than age 30 and older than 50 years did not
differ in their rates of increase in SBP and DBP.
The relationships between physical activity and blood pressure and other
biological CVD risk factors that are related to blood pressure (e.g., age, BMI, and
smoking) were assessed using mixed random effects models. Table 9 displays the
results of mixed models using SBP and DBP as the dependent variables, and
physical activity levels, time since first exam, and the interaction of the two as the
independent variables. Cross-sectionally, mean SBP significantly varied by physical
activity level (p<0.001). Relative to subjects active 5 days/week, those who were
active 1 to 3 days per week had significantly lower SBP. Sedentary subjects did not
differ from the highly active subjects in the cross-sectional estimate of the mean SBP
(p=0.56). The average rate of change in SBP in subjects active 5 days/week was
0.31 mm Hg per year (p<0.001). The rate of change in SBP significantly varied by
physical activity level (p<0.001), with subjects active less then 4 days per week all
having significantly higher rates of increase in SBP than the highly active group. In
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 8. Longitudinal analysis of the influence of age on physical activity and blood pressure: Dependent variable = physical
activity, SBP, DBP; independent variables = age, years since first examination, and interaction of age * years since
first examination.
Physical activity SBP DBP
Estimate p-value Estimate p-value Estimate p-value
Intercept 3.29 <. 001 126.6 <.001 81.78 <.001
Age (groups)3 <.001b <.001b <.001b
<30 years 0.50 <. 001 -0.67 0.45 -5.26 <.001
3 0 -4 0 years 0.22 0.006 -6.01 <.001 -6.31 <.001
40 ~ 50 years 0.05 0.56 -5.22 <.001 -3.15 <.001
>= 50 years 0 0 0
Years 0.003 0.67b 0.32 <.001b 0.21 <.001b
Years * Agea 0.08b < ,001b <.001b
< 30 years -0.003 0.06 -0.14 0.35 0.16 0.10
3 0 -4 0 years 0.007 0.44 0.61 <.001 0.53 <.001
4 0 -5 0 years 0.002 0.77 0.42 <.001 0.22 0.001
>= 50 years 0 0 0
• A mixed random effects linear model was used.
• a Age >=50 years is reference category.
• b p-value from Type 3 tests of fixed effects.
to
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 9. Longitudinal analysis of the relationship between levels of physical activity and blood pressure: Dependent variables =
SBP, DBP; independent variables = physical activity, years since first exam, interaction term of activity * years since
first exam.
SBP DBP
Estimate p-value Estimate p-value
Intercept 124.70 <.001 76.87 <.001
Physical activity (levels)a
A
o
o
CP
0.004b
Sedentary -0.67 0.56 1.73 0.02
1 day / week -2.21 0.006 -0.15 0.78
2 days / week -2.73 0.0003 0.22 0.65
3 days / week -1.98 0.007 -0.39 0.42
4 days / week 0.76 0.40 0.79 0.18
5 days / week 0 0
Years 0.31 <.001b 0.59 <.001b
Years* Physical activity (levels)a <.001b <.001b
Sedentary 0.27 0.047 -0.06 0.46
1 day / week 0.59 <.001 0.15 0.04
2 days / week 0.48 <.001 0.08 0.25
3 days / week 0.30 0.003 -0.02 0.72
4 days / week -0.02 0.86 -0.16 0.04
5 days / week 0 0
• A mixed random effects linear model was used.
• a The highest level of physical activity (5 days / week) is reference category.
• b p-value from Type 3 tests of fixed effects.
to
C\
contrast, cross-sectional results for DBP as the dependent variable showed only a
significant difference in the mean DBP between highly active (5 days/week) and
sedentary subjects, with the sedentary subjects having a higher mean DBP (p=0.02).
Although the rates of change in DBP significantly varied by physical activity level
(p<0.001), there was no clear pattern of differences by physical activity level.
The relationship between change in longitudinal physical activity and blood
pressure was further evaluated in a multivariate model, in which several potential
confounding parameters were included (Table 10). After adjusting for age, BMI and
smoking status, the inverse cross-sectional relationship noted between physical
activity and systolic blood pressure was still statistically significant (adjusted p-value
< 0.001 for cross-sectional differences in SBP by physical activity level). The
adjusted association between physical activity level and rates of change in SBP and
DBP (p < 0.001 for both) also remained highly significant. Relative to those active 5
days/week, subjects active 1-3 days weekly had significantly higher rates of annual
increase in SBP. As with the unadjusted model, there was no clear pattern of
differences in DBP change rates by physical activity level. In this multivariate
model, age and smoking were significantly associated with higher rates of increase in
SBP (p < 0.001 for both). BMI was positively associated with higher SBP and DBP
cross-sectionally, but was not associated with differential rates of change in SBP or
DBP. Smoking (p=0.0005) was associated with higher rates of increase in diastolic
blood pressure, while age was associated with lower rates of change in DBP. BMI
was not associated with differential rates of change in DBP.
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 10. Longitudinal analysis of the relationship between the levels of physical activity and blood pressures with adjustment
for other CVD risk factors: Dependent variable = SBP, DPB; independent variables = physical activity, years since first
exam, and the interaction of activity * years since first exam.
SBP DBP
Estimate p-value Estimate p-value
Intercept 105.25 < .0001 51.89 <.0001
Physical activity (levels)3 <.0001b <.0001b
Sedentary -0.21 0.85 1.13 0.13
1 day/ week -1.88 0.02 -1.11 0.04
2 days / week -2.41 0.001 -0.70 0.16
3 days / week -1.79 0.01 -1.07 0.02
4 days / week 0.82 0.35 0.54 0.34
5 days / week 0 0
Years -0.39 0.37b 0.61 0.003b
Years * Physical activity (levels)3 <.0001b <.0001b
Sedentary 0.07 0.63 -0.10 0.25
1 day/ week 0.43 <.0001 0.16 0.03
2 days / week 0.34 0.001 0.10 0.14
3 days / week 0.20 0.04 -0.003 0.97
4 days / week -0.08 0.52 -0.15 0.04
5 days / week 0 0
Covariate variables
Age -0.13 <.0001b 0.16 < ,0001b
Years * Age 0.02 <.0001b -0.005 0.03b
BMI 1.07 <.0001b 0.89 < .000 lb
Years * BMI -0.005 0.62b -0.01 0.18b
K)
00
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 10. (Continued)
SBP DBP
Estimate p-value Estimate p-value
Covariate variables
Smoking status0 -1.75 < ,0001b -1.05 < ,0001b
Years * Smoking status0 0.13 <.0001b 0.07 0.0005b
• A mixed random effects linear model was used.
• a The highest level of physical activity (5 days / week) is reference category.
• b p-value from Type 3 tests of fixed effects.
• c Current smoking status is reference category.
to
VO
2. Longitudinal analyses of physical fitness and the association with blood
pressure
To examine the longitudinal relationship between physical fitness and blood
pressure, physical fitness was recoded as a categorical parameter with four levels
roughly divided by its quartile distribution. The results of univariate random effects
mixed models with systolic and diastolic blood pressure as dependent variables are
given in Table 11. Cross-sectionally, lower levels of physical fitness were
significantly associated with lower mean SBP (p < 0.0003). The annual rate of
change in SBP in the highest physical fitness level was 0.44 mm Hg per year
(p<0.001). Physical fitness was significantly associated with the rates of yearly SBP
change (p < 0.001), such that subjects at lower levels of physical fitness had higher
mean yearly increases in SBP than did the highest physical fitness group. Similarly,
cross-sectional associations showed that physical fitness was positively associated
with mean DBP (p < 0.001), with mean DBP levels lower in the lower physical
fitness groups relative to the highest quartile of physical fitness. Physical fitness was
also significantly associated with the rates of longitudinal DBP change (p < 0.001),
with the mean DPB change rates higher in subjects with lower levels of physical
fitness.
In a multivariate random effects regression model, the potential confounding
factors were included in the model, and the longitudinal relationships between
physical fitness and blood pressure were assessed (Table 12). After adjusting for
age, BMI and smoking status, the cross-sectional associations noted in unadjusted
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 11. Longitudinal analysis of the relationship between the levels of physical fitness and blood pressure: Dependent variable
= SBP, DBP; independent variables = physical fitness level, years since first exam, and the interaction of the two.
SBP DBP
Estimate p-value Estimate p-value
Intercept 124.36 <.001 77.45 <.001
Physical fitness (levels)3 0.0003b 0.009b
< 135 watts -1.22 0.038 -0.28 0.46
135 ~ 160 watts -2.30 <.001 -1.08 0.003
160 ~ 185 watts -1.92 0.001 -0.81 0.03
>185 watts 0 0
Years 0.44 < ,001b 0.49 < ,001b
Years* Physical fitness (levels)3 <.001b <.001b
< 135 watts 0.28 <.001 0.20 <.001
135 ~ 160 watts 0.36 <.001 0.20 <.001
160- 185 watts 0.22 0.001 0.13 0.003
> 185 watts 0 0
• A mixed random effects linear model
• a The highest level of physical fitness (> 185 watts is reference category.
• b p-value from Type 3 tests of fixed effects.
u >
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 12. Longitudinal analysis of the relationship between the levels of physical fitness and blood pressure with adjustment for
other CVD risk factors: Dependent variable = SBP, DBP; independent variables = physical fitness level, years since
first exam, and the interaction of the two.
SBP DBP
Estimate p-value Estimate p-value
Intercept 105.98 <.0001 52.01 <.0001
Physical fitness (levels)2
o
o
©
0.007b
< 135 watts -0.42 0.47 0.15 0.69
135 ~ 160 watts -1.77 0.002 -0.83 0.02
160- 185 watts -1.63 0.004 -0.63 0.09
>185 watts 0 0
Years -0.48 0.22b 0.48 0.0005b
Years * Physical fitness (levels)2 < ,0001b 0.0005b
< 135 watts 0.21 0.003 0.15 0.002
135 - 160 watts 0.31 < .0001 0.17 <.0001
160- 185 watts 0.19 0.004 0.11 0.01
> 185 watts 0 0
Covariate variables
Age -0.15 < .000l b 0.15 <.0001b
Years * Age 0.02 <.0001b -0.005 0.02b
BMI 1.05 < .000l b 0.88 < ,0001b
Years * BMI 0.001 0.89b -0.004 0.47b
Smoking status0 -1.76 < .000l b -1.04 <.0001b
Years * Smoking status0 0.13 < .000 l b 0.06 0.002b
• A mixed random effects linear model was used.
• a The highest level of physical fitness (> 185 watts) is reference category.
• b p-value from Type 3 tests of fixed effect.
• c Current smoking status is reference category.
t o
models between physical fitness and blood pressure remained significant (p=0.001
for SBP, p=0.007 for DBP). The rates of change in both SBP (p<0.001) and DBP
(p=0.0005) significantly varied by physical fitness level in these covariate-adjusted
models. Relative to the highest quartile of physical fitness, rates of SBP and DBP
change were significantly higher as physical fitness decreased (SBP: p<0.0001,
DBP: p=0.0005). In these multivariate models, the rates of change in SBP were
significantly greater at higher ages (p<0.0001), but rates of change in DBP were
significantly lower at higher ages (p=0.024). Rates of change in SBP and DBP were
significantly higher in smokers than non-msokers (SBP: pO.OOOl, DBP: p=0.002).
BMI was positively associated with higher SBP and DBP cross-sectionally (both
p<0.001), but BMI was not associated with differential rates of change in either SBP
or DBP (SBP: p=0.89, DBP: p=0.47).
3. Evaluation of possible modifying effects of physical fitness level and follow-up
examination period on the association between physical activity and blood pressure
The changes in physical fitness and physical activity were compared between
first measurement and last measurement using the paired t-test. Compared to the
first measurement, the mean physical fitness of the last measurement was
significantly higher (168.53 ± 39.57 vs. 152.34 ± 35.34, p < 0.001) with a 10.6%
increase, whereas the mean physical activity was significantly lower (3.37 ± 1.28 vs.
3.56 ± 1.21, p < 0.001) with a 5.3% decrease. In addition, within each level of
physical activity, physical fitness measured at the last examination was higher than
that measured at the first examination (Figure 1). The ratios of the last measured to
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 1. Comparison of physical fitness between
first and last measurement
T otal
£
.!5 o
tS " S
- I
£
P h
■ g g p a
El Last measurement
0 First measurement
w s s s s fffs s ffs fs s s s s s /ffjw s s s s s jr s ffs jv fs s s s s ffffjr s /
H
f j r S S S S S S S S S S S f S S S S ffS S f fj r S S S S S S fS S S j r S S S S S S S j
50 100 150 200
Physical fitness (watts/min)
250
Figure 2. Longitudinal change in physical fitness between
first and last measurement
0 0
d
d
c j
• M
5 0
js*
a
%
o
c
V
S
a >
L .
9
V i
C Q
C L »
a
ts
1.20
1.15
1.10
1.05
1.00-
S N
/ r V
0 1 2 3 4 5
Physical activity (days/week)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the first measured physical fitness generally increased with the increasing levels of
physical activity (Figure 2). Change in physical fitness was not correlated with
baseline physical activity (first measurement) (Pearson’s correlation coefficient p = -
0.02, p = 0.37) whereas change in physical fitness was positively correlated with the
last measured physical activity (Pearson’s correlation coefficient p = 0.22, p <
0.001).
The modifying effects of physical fitness on the relationship between
longitudinal change in physical activity and blood pressure among different levels of
physical fitness are summarized in Tables 13 and 14. Cross-sectionally, the mean
SBP significantly varied by activity level within all physical fitness groups (Table
13, p=0.03, 0.02, 0.047, and 0.015 respectively from low to high level of physical
fitness). In general, the mean SBP was lower in less physically active categories
(relative to the referent group of activity 5 days/week). However, in the highest
quartile of physical fitness, the mean SBP was higher in the groups with low physical
activity (sedentary and 1 day/week). Cross-sectionally, levels of DPB significantly
varied by physical activity only in the lower two quartile subgroups of physical
fitness (Table 14, p=0.04 and 0.007, respectively). In these two subgroups, the
cross-sectional mean DBP tended to be higher in the less active groups relative to the
most highly active group (5 days/week). In general, the rates of change in SBP and
DBP significantly varied by physical activity level throughout the range of physical
fitness (Tables 13 and 14). The mean rates of change in SBP tended to be higher in
less physically active groups (Table 13). Associations between physical activity and
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 13. Longitudinal analysis of the univariate relationship between physical activity and SBP by physical fitness level:
Dependent variable = SBP, independent variables = physical activity, years since first exam, and the interaction
of the two.
Physical fitness
( < 135 watts )
Physical fitness
( 135-160 watts)
Physical fitness
( 160-185 watts )
Physical fitness
(>= 185 watts)
SBP SBP SBP SBP
Estimate p-value Estimate p-value Estimate p-value Estimate p-value
Intercept 125.33 <.001 124.63 <.001 124.17 <.001 125.07 <.001
Physical activity (levels)3 0.03b 0.02b 0.047b 0.015b
Sedentary -0.48 0.85 -2.28 0.37 1.51 0.58 2.91 0.35
1 day / week -3.45 0.13 -2.96 0.11 -2.44 0.17 3.85 0.04
2 days / week -3.28 0.14 -3.31 0.06 -3.41 0.04 -2.21 0.15
3 days / week -2.11 0.34 -3.63 0.04 -1.80 0.24 -1.21 0.31
4 days / week 2.19 0.43 1.24 0.56 0.56 0.76 0.30 0.83
5 days / week 0 0 0 0
Years 0.58 0.17 0.17 0.53 0.39 0.046 0.13 0.31
Years * Physical activity (levels)3 0.06b 0.003b 0.04b 0.05b
Sedentary -0.06 0.90 0.44 0.20 0.06 0.84 -0.01 0.98
1 day / week 0.40 0.36 0.69 0.02 0.44 0.06 0.03 0.90
2 days / week 0.18 0.68 0.68 0.02 0.30 0.17 0.47 0.007
3 days / week -0.04 0.92 0.59 0.04 0.15 0.48 0.27 0.07
4 days / week -0.10 0.84 -0.07 0.83 -0.17 0.49 0.09 0.61
5 days / week 0 0 0 0
• A mixed random effects linear model was used.
• aThe highest level of physical activity (5 days / week) is reference category.
• b p-value from Type 3 tests of fixed effects.
U >
Os
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 14. Longitudinal analysis of the univariate relationship between physical activity and DBP by physical fitness levels:
Dependent variable = DBP, independent variables = physical activity, years since first exam, and the interaction of
the two.
Physical fitness
(< 135 watts )
Physical fitness
( 135-160 watts)
Physical fitness
( 160-185 watts)
Physical fitness
(>= 185 watts )
DBP DBP DBP DBP
Estimate p-value Estimate p-value Estimate p-value Estimate p-value
Intercept 76.74 <.001 75.08 <.001 76.60 <.001 77.72 <.001
Physical activity (levels)3 0.040b 0.007b 0.5 lb 0.47b
Sedentary 2.41 0.16 5.00 0.002 1.79 0.32 1.55 0.44
1 day / week 0.02 0.99 1.95 0.10 -0.98 0.40 1.14 0.34
2 days / week 0.78 0.60 1.24 0.27 0.11 0.92 -0.58 0.55
3 days / week -0.03 0.98 0.61 0.58 0.15 0.88 -0.74 0.34
4 days / week 3.26 0.08 2.27 0.10 0.73 0.55 -0.39 0.67
5 days / week 0 0 0 0
Years 0.84 <.001 0.59 0.001 0.58 <.001 0.40 <.001
Years * Physical activity (levels)3 0.04b 0.04b 0.003b 0.71b
Sedentary -0.36 0.24 -0.32 0.14 -0.02 0.92 0.08 0.71
1 day/week -0.04 0.90 0.10 0.61 0.22 0.16 -0.10 0.45
2 days / week -0.14 0.63 0.12 0.50 0.04 0.77 0.11 0.34
3 days / week -0.21 0.46 0.03 0.87 -0.10 0.47 0.03 0.76
4 days / week -0.57 0.09 -0.08 0.69 -0.31 0.06 -0.01 0.93
5 days / week 0 0 0 0
• A mixed random effects linear model was used.
• a The highest level of physical activity (5 days / week) is reference category.
• b p-value from Type 3 tests of fixed effects.
DBP rates of change showed no clear pattern. At the highest level of physical
fitness, rates of DBP change did not significantly differ among the physical activity
levels (Table 14).
The possible modifying effects of time course on the association between
physical activity and blood pressure were assessed. Starting from the first
examination, the relationships regarding physical activity on systolic blood pressure
were analyzed by different follow-up periods (short term defined as years 1-2,
medium term as years 3-5, and long term as years 6-10 and years 11-20) (Table 15).
Cross-sectionally, the mean SBP significantly varied by level of physical activity in
all follow-up periods (all p < 0.001). Among every follow-up period, the general
quadratic trends in the mean SBP observed in the original cross-sectional analysis
using first examination data only was seen. Subjects reporting mild and moderate
physical activity (exercise 1 to 3 days per week) had lower mean SBP than subjects
exercising 5 days/week. Except for the early follow-up period (1-2 years), the mean
rate of change in SBP significantly varied by physical activity level, with less active
subjects tending to show higher average rates of increase of SBP than the highly
active group. Like the cross-sectional analyses, the sedentary group did not tend to
show different rates of change in SBP relative to the highly active group.
Similar analyses were done for diastolic blood pressure (Table 16). Cross-
sectionally, the mean DBP significantly varied by physical activity level only in the
second (p=0.01) and third (p=0.009) follow-up periods. Rates of DBP change
significantly varied by physical activity level in all but the shortest follow-up period.
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 15. Longitudinal analysis of the relationship between physical activity and SBP by follow-up measurement periods:
Dependent variable = SBP; independent variables = physical activity, years since first exam, physical fitness, and the
interaction of the physical activity/fitness with years since first exam.
Measurement period
(1-2 years)
Measurement period
(3-5 years)
Measurement period
(6-10 years)
Measurement period
(11-20 years)
SBP SBP SBP SBP
Estimate p-value Estimate p-value Estimate p-value Estimate p-value
Intercept 127.35 <.001 125.26 <.0001 127.17 <.0001 126.62 < .0001
Physical activity (levels)3 0.002b < ,0001b < ,0001b < .0001b
Sedentary 0.86 0.61 0.24 0.87 0.04 0.98 -0.55 0.72
1 day/ week -3.88 0.01 -2.84 0.007 -3.68 0.0004 -3.49 0.0001
2 days / week -4.92 0.001 -3.45 0.0004 -4.65 < .0001 -4.60 < .0001
3 days / week -2.79 0.02 -2.01 0.03 -2.50 0.006 -2.77 0.003
4 days / week 0.51 0.68 1.32 0.25 0.81 0.48 0.32 0.78
5 days / week 0 0 0 0
Years 0.82 0.75 0.34 0.55 0.34 0.12 0.33 0.05
Years *Physical activity (levels)3 0.52b 0.01b <.0001b < ,0001b
Sedentary 0.06 0.98 -0.29 0.60 0.03 0.91 0.24 0.13
1 day / week 1.97 0.32 0.61 0.12 0.94 <.0001 0.49 < .0001
2 days / week 1.92 0.33 0.68 0.07 0.65 0.0001 0.54 < .0001
3 days / week 0.28 0.87 0.05 0.90 0.34 0.03 0.32 0.004
4 days / week 0.13 0.94 -0.37 0.40 -0.09 0.63 0.03 0.84
5 days / week 0 0 0 0
Physical fitness -0.01 0.28b -0.001 0.84b -0.008 0.26b -0.005 0.5 lb
Years * Physical fitness -0.01 0.49b -0.002 0.48b -0.0003 0.80b -0.001 0.40b
• A mixed random effects linear model was used.
• a The highest level of physical activity (5 days / week) is reference category.
• b p-value from Type 3 tests of fixed effects.
V O
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 16. Longitudinal analysis of the relationship between physical activity and DBP by measurement periods: Dependent
variable = DPB; independent variables = physical activity, years since first exam, physical fitness, and the
interaction of activity/fitness with years since first exam.
Measurement period
(1-2 years)
Measurement period
(3-5 years)
Measurement period
(6-10 years)
Measurement period
(11-20 years)
DBP DBP DBP DBP
Estimate p-value Estimate p-value Estimate p-value Estimate p-value
Intercept 78.14 <.001 77.65 <.001 78.43 <.001 78.02 <.001
Physical activity (levels)3 0.10b 0.01b 0.009b 0.07b
Sedentary 1.83 0.13 1.36 0.19 1.28 0.20 1.21 0.25
1 day / week -0.39 0.61 -0.31 0.66 -0.87 0.21 -0.45 0.53
2 days / week -0.04 0.96 0.30 0.64 -0.14 0.83 -0.15 0.82
3 days / week -0.23 0.73 0.03 0.97 -0.42 0.49 -0.38 0.55
4 days / week 1.62 0.08 2.02 0.01 1.38 0.07 1.26 0.11
5 days / week 0 0 0 0
Years 1.16 0.51 0.71 0.06 0.78 <.001 0.62 <.001
Years * Physical activity (levels)3 0.29b 0.002b < ,001b 0.005b
Sedentary 1.22 0.46 0.02 0.96 -0.12 0.44 0.000 1 .000
1 day / week 1.43 0.29 0.23 0.38 0.22 0.07 0.14 0.10
2 days / week -0.26 0.84 -0.02 0.95 0.06 0.58 0.10 0.18
3 days / week -0.28 0.82 -0.23 0.32 -0.05 0.65 -0.001 0.99
4 days / week -0.38 0.78 -0.85 0.005 -0.31 0.01 -0.11 0.22
5 days / week 0 0 0
Physical fitness -0.010 0.09b -0.008 0.10b -0.010 0.04b -0.008 0.1 l b
Years * Physical fitness 0.0005 0.95b -0.0003 0.88b -0.0002 0.73b -0.0005 0.3 lb
• A mixed random effects linear model was used.
• a The highest level of physical activity (5 days / week) is reference category.
• b p-value from Type 3 tests of fixed effects.
o
There was no clear pattern of association of physical activity with cross-sectional
DBP means or DBP change rates.
In both Tables 15 and 16, analyses were also adjusted for physical fitness and
the influence of physical fitness on blood pressure change rates (years*physical
fitness). Interestingly, the associations of physical fitness with blood pressure
change rates noted in earlier analyses were no longer significant after adjustment for
physical activity.
DISCUSSION
The results of this study indicate significant relationships among physical
activity, physical fitness, and systolic (SBP) and diastolic (DBP) blood pressure in
this Los Angeles County firefighter cohort, in which the variables of interest were
repeatedly measured over a period of 20 years. Physical activity and physical fitness
were positively correlated. Physical activity and physical fitness were negatively
related to SBP and DBP at any time point (cross-sectionally) and the annual rates of
increase in blood pressure were reduced with higher levels of physical activity and
fitness. In the cross-sectional analysis using data from the baseline examination,
there was an anomalous quadratic trend in blood pressure, such that blood pressure
decreased and then increased as physical activity and fitness increased. In the final
multivariate analysis, the relationships between physical activity and fitness and
blood pressure reduction remained significant. In particular, moderately intense
physical activity was more beneficial for lowering SBP.
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Many literature reviews and meta-analyses suggest that regular exercise (i.e.,
walking, jogging and cycling) lowers the resting blood pressure (van Baak 1998;
Lesniak and Dubbert 2001), and that improved physical fitness along with habitual
physical activity are associated with blood pressure reduction (Kokkinos and
Papademetriou 2000). However, the evidence to support this notion is primarily
derived from observational and cross-sectional studies, and the strength and
consistency of the evidence vary from one study to another study due to the
differences and limitations of these study designs. The findings from the meta
analyses across studies indicate that both SBP and DBP can be reduced (3-10 mmHg
SBP, 3-8 mmHg DBP) and an increased DBP with exercise was occasionally found
(van Baak 1998).
In the present study, the analysis of the relationships between physical
activity and SBP and DBP from cross-sectional data using the first measure (baseline
assessment) demonstrated inconsistent results with the findings from the longitudinal
analyses. At the baseline examination, SBP was inversely linearly related to both
physical activity and physical fitness. However, there was also a highly significant
positive quadratic trend such that SBP decreased with moderate levels of
activity/fitness, then increased at the highest levels of activity/fitness (Tables 3 and
4). DBP at the baseline examination showed a positive linear trend with physical
activity, and the same inverse linear and positive quadratic relationship with physical
fitness (as observed for SBP). On longitudinal analyses using the entire observation
period, subjects with higher levels of physical activity and fitness showed lower rates
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
of increase of SBP, with no clear pattern of association with rates of change of DBP.
The reason why physical activity and fitness showed this quadratic pattern of
initially decreasing then increasing SBP on cross-sectional analysis is not clear. In a
cross-sectional analysis, cohort effects and selection biases may bias analyses of
relationships. For example, blood pressure may have been initially higher in subjects
at the highest levels of physical activity and fitness because knowledge of their high
blood pressure led them to exercise more. Notably, this quadratic pattern was no
longer apparent on cross-sectional analyses using data from the last examination
(Table 5). At the last examination, the inverse relationship between physical
activity/fitness and blood pressure was apparent, with the lowest mean blood
pressure seen in subjects at the highest levels of activity/fitness. A longitudinal
observational study is the better experimental design, because both lifestyle (e.g.,
physical activity behavior) and other chronic disease risk factors (e.g., blood
pressure) are measured repeatedly over time, and associations on changes in these
variables will not reflect initial selection biases.
A longitudinal design in the present study was its primary advantage and its
main purpose was to investigate the longitudinal relationships among physical
activity and physical fitness as predictor variables and SBP and DBP as outcome
variables, which are “intermediate” outcome measures for cardiovascular diseases.
Therefore, a longitudinal linear regression analysis was used, which is extensively
described elsewhere (Lefevre et al. 1999; Twisk et al. 2001). The longitudinal
regression coefficients were estimated with a random effects model, which is
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
different from the standard linear regression model. The standard linear model is
only good for balanced longitudinal data, which requires that all subjects should have
the same number of repeated measures and the repeated measures should be assessed
at the same time points for all subjects. Consequently, only the data of the subjects
with all measurements available are used in the calculations. On the other hand, a
mixed linear model uses all available data (not only the “complete cases”) in the
analysis, and all the measurements can be viewed as being measured on a continuous
rather than discrete time scale. The use of random effects in the mixed linear
analysis allows modeling of covariances as continuous functions of time and
inclusion of time-varying covariates in the mean structure, which is not possible in a
standard linear regression model (Verbeke and Molenberghs 2000).
We observed that the longitudinal changes in physical activity and physical
fitness were inversely related to rates of change in SBP. Higher levels of physical
fitness, but not physical activity, were longitudinally inversely related to rates of
change in DBP. Relative to 5 days/week, mild to moderate physical activity (2-4
days/week) was significantly associated with higher rates of change in SBP in both
univariate and multivariate analysis (Tables 9 and 10). The annual rate of increase in
SBP was estimated to be approximately 0.3-0.6 mmHg higher relative to 5
days/week of activity. There was no clear pattern of association between DBP and
the levels of physical activity. The inverse relationship between longitudinal change
in fitness and SBP rates of change was significant in both univariate and multivariate
analyses (Tables 11 and 12) and lower levels of fitness were associated with rates of
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
SBP increase that were 0.2-0.36 mmHg/year higher than the highest quartile of
fitness. In addition, physical fitness was inversely related to DBP change rates
(0.13-0.2 mmHg/year higher in lower levels of fitness compared to the highest level).
The relatively small blood pressure reductions found in this study are in
agreement with the literature. Physical activity may induce a greater reduction in
blood pressure among persons with hypertension compared to normotensive controls
(van baak 1998; Kokkinos and Papademetriou 2000). Because the subjects in our
cohort were firefighters and most of them were healthy persons (less than 5% of the
subjects were hypertensive), the levels of blood pressure reduction detected in this
study were likely lower than might have been seen if the cohort had included more
hypertensive subjects. Several longitudinal studies [i.e., the Northern Ireland Young
Heart Study (Boreham et al. 2002), the Danish Youth and Sports Study (Hasselstrom
et al. 2002), the Muscatine Study (Janz et al. 2002), the Leuven Longitudinal Study
on Lifestyle, Fitness and Health (Lefevre et al. 2002), and the Amsterdam Growth
and Health Longitudinal Study (Twisk et al. 2002)] showed no significant
relationships among changes in physical activity and physical fitness on either
systolic or diastolic blood pressure. One explanation for this difference is that most
of these cited studies evaluated physical activity/fitness in adolescence in relation to
blood pressure measured as young adults. The present study evaluated these
relationships in adults with a mean age 33 years at the first assessment and 44 at the
end of the follow-up. Thus, the association between physical activity/fitness and
blood pressure may be age-related and/or related to initial levels of blood pressure.
45
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
It is also possible that subjects in these previous studies were healthier than subjects
in our study. Finally, the small sample size in some of these studies could have
reduced the power to detect true associations.
In the multivariate longitudinal analysis, body mass index was significantly
positively related to both SBP and DBP at any time point (cross-sectionally), but was
not associated with rates of change in SBP or DBP. In contrast, smoking was
significantly associated with lower SBP and DBP at any time point and with higher
rates of increase in SBP and DBP, relative to non-smokers. The association with
age varied for systolic and diastolic blood pressure. Age was inversely associated
with SBP at any time point, but positively associated with rates of change in SBP.
Age was positively associated with DBP at any time point, but inversely associated
with rates of change in DBP. Therefore, the relationships between change in
physical activity and physical fitness and blood pressure were possibly confounded
by these factors. Exercises in a loss of body fat, leading to a redistribution of fat
store and weight loss (lower BMI), and is subsequently associated with a
concomitant lowered blood pressure (Borhani 1996). Our study also showed that the
levels of physical activity were highly negatively associated with BMI (cross-
sectionally), suggesting that lowering BMI by physical activity is one of the
mechanisms for the association of physical activity with reduced blood pressure.
The cross-sectional negative relationship between smoking and SBP and DBP seems
contradictory, but this phenomenon has also been reported in other studies (Handa et
al. 1990; Twisk et al. 1997). The acute response to smoking is to elevate blood
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
pressure, while the long-term response to smoking is to lower blood pressure
(Kiowski et al. 1994). While the explanation for this finding is not clear, one
possible explanation is that smoking may have a vascular relaxation effect, which
could lower blood pressure.
Physical activity and physical fitness are highly correlated. A meta-analysis
suggested that the association of physical fitness with cardiovascular diseases was
significantly different from physical activity, and the cardiovascular disease risk
reduction was significantly greater for physical fitness than physical activity
(Williams 2001). Different levels of physical fitness between individuals play an
important role in energy expenditure, and physical activities with the same energy
expenditure are less physically stressful for individuals with high levels of physical
fitness. It has been argued that physical activity must reach a certain level (a
threshold) to improve one's physical fitness and achieve a protective effect (Kannel
et al. 1986; Lesniak and Dubbert 2001). To evaluate how the longitudinal
relationship between physical activity and the reduction of blood pressure might be
modified by level of physical fitness, an additional analysis was conducted. At each
level of physical fitness, the association between change in physical activity and
blood pressure was assessed. An inverse relationship between physical activity and
SBP consistently existed at each level of physical fitness.
To investigate whether the effect of physical activity on blood pressure was
long lasting or short-term, the longitudinal association between physical activity and
blood pressure was assessed during different follow-up time periods. Mild to
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
moderate physical activity (exercise 1 to 3 days per week) was consistently
negatively associated with systolic blood pressure in all follow-up periods. Taken
together, these results indicate that physical activity has both a short- and long-term
beneficial association with systolic blood pressure.
In summary, we report that physical activity and physical fitness are related
to reduced rates of SBP change in longitudinal observations. The positive benefits of
physical activity on physical fitness are apparent throughout the range of achieved
physical fitness, and show both a short- and long-term association with blood
pressure. In addition, this study indicates that physical activity and physical fitness
appear to be more strongly associated with SBP than that for DBP, and physical
fitness may be more important in defining this health benefit. These findings also
provide evidence to support the recommendation of the U.S. Centers for Disease
Control and Prevention, which is that moderate-intensity physical activity (i.e., 30 to
60 minutes of brisk walking), should be done on most days of the week (Pate et al.
1995). While this study used a self-report of the number of days per week of
vigorous exercise, a person’s total daily energy expenditure may be a more
informative parameter for studying physical activity associations. To quantify
physical activity as energy expenditure, the frequency, duration and intensity of each
physical activity must be measured. Therefore, measuring physical activity in terms
of energy expenditure rather than general categories of activity may be difficult to
achieve in large epidemiologic studies. In addition, this is not a population-based
observational study. Results of the current study can be generalized to generally
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
healthy employed males. More work needs to be done to determine the optimal type
of exercise and dose for various population subgroups.
REFERENCES
Blair S. N. and Jackson A. S. (2001) Physical fitness and activity as separate heart
disease risk factors: a meta-analysis. M ed Sci Sports Exerc 33, 762-764.
Blair S. N., Cheng Y. and Holder J. S. (2001) Is physical activity or physical fitness
more important in defining health benefits? M ed Sci Sports Exerc 33, S379-
399; discussion S419-320.
Boreham C., Twisk J., Neville C., Savage M., Murray L. and Gallagher A. (2002)
Associations between physical fitness and activity patterns during
adolescence and cardiovascular risk factors in young adulthood: the Northern
Ireland Young Hearts Project. IntJSports M ed 23 Suppl 1, S22-26.
Borhani N. O. (1996) Significance of physical activity for prevention and control of
hypertension. JHum Hypertens 10 Suppl 2, S7-11.
Church T. S., Kampert J. B., Gibbons L. W., Barlow C. E. and Blair S. N. (2001)
Usefulness of cardiorespiratory fitness as a predictor of all-cause and
cardiovascular disease mortality in men with systemic hypertension. Am J
Cardiol 88, 651-656.
Erikssen G. (2001) Physical fitness and changes in mortality: the survival of the
fittest. Sports M ed 31, 571-576.
Guidotti T. L. (1992) Human factors in firefighting: ergonomic-, cardiopulmonary-,
and psychogenic stress-related issues. Int Arch Occup Environ Health 64, 1-
12.
Handa K., Tanaka H., Shindo M., Kono S., Sasaki J. and Arakawa K. (1990)
Relationship of cigarette smoking to blood pressure and serum lipids.
Atherosclerosis 84, 189-193.
Hasselstrom H., Hansen S. E., Froberg K. and Andersen L. B. (2002) Physical fitness
and physical activity during adolescence as predictors of cardiovascular
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
disease risk in young adulthood. Danish Youth and Sports Study. An eight-
year follow-up study. IntJSports M ed 23 Suppl 1, S27-31.
Horowitz M. R. and Montgomery D. L. (1993) Physiological profile of fire fighters
compared to norms for the Canadian population. Can J Public Health 84, 50-
52.
Janz K. F., Dawson J. D. and Mahoney L. T. (2002) Increases in physical fitness
during childhood improve cardiovascular health during adolescence: the
Muscatine Study. Int J Sports M ed 23 Suppl 1, S15-21.
Kannel W. B., Belanger A., DAgostino R. and Israel I. (1986) Physical activity and
physical demand on the job and risk of cardiovascular disease and death: the
Framingham Study. Am Heart J 112, 820-825.
Katzmarzyk P. T., Malina R. M. and Bouchard C. (1999) Physical activity, physical
fitness, and coronary heart disease risk factors in youth: the Quebec Family
Study. Prev M ed 29, 555-562.
Kavanagh T. (2001) Exercise in the primary prevention of coronary artery disease.
Can J Cardiol 1 7 ,155-161.
Kiowski W., Linder L., Stoschitzky K., Pfisterer M., Burckhardt D., Burkart F. and
Buhler F. R. (1994) Diminished vascular response to inhibition of
endothelium-derived nitric oxide and enhanced vasoconstriction to
exogenously administered endothelin-1 in clinically healthy smokers.
Circulation 90,27-34.
Kohl H. W., 3rd (2001) Physical activity and cardiovascular disease: evidence for a
dose response. M ed Sci Sports Exerc 33, S472-483; discussion S493-474.
Kokkinos P. F. and Papademetriou Y. (2000) Exercise and hypertension. Coron
Artery Dis 11, 99-102.
Lee I. M., Hsieh C. C. and Paffenbarger R. S., Jr. (1995) Exercise intensity and
longevity in men. The Harvard Alumni Health Study. Jama 273, 1179-1184.
Lefevre J., Philippaerts R., Delvaux K., Thomis M., Claessens A. L., Lysens R.,
Renson R., Vanden Eynde B., Vanreusel B. and Beunen G. (2002) Relation
between cardiovascular risk factors at adult age, and physical activity during
youth and adulthood: the Leuven Longitudinal Study on Lifestyle, Fitness
and Health. Int J Sports M ed 23 Suppl 1, S32-38.
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Lesniak K. T. and Dubbert P. M. (2001) Exercise and hypertension. Curr Opin
Cardiol 16, 356-359.
Oja P. (2001) Dose response between total volume of physical activity and health
and fitness. M ed Sci Sports Exerc 33, S428-437; discussion S452-423.
Pate R. R., Pratt M., Blair S. N., Haskell W. L., Macera C. A., Bouchard C., Buchner
D., Ettinger W., Heath G. W., King A. C. and et al. (1995) Physical activity
and public health. A recommendation from the Centers for Disease Control
and Prevention and the American College of Sports Medicine. Jama 273,
402-407.
Powell K. E. and Blair S. N. (1994) The public health burdens of sedentary living
habits: theoretical but realistic estimates. M ed Sci Sports Exerc 26, 851-856.
Powell K. E., Thompson P. D., Caspersen C. J. and Kendrick J. S. (1987) Physical
activity and the incidence of coronary heart disease. Annu Rev Public Health
8, 253-287.
Rankinen T. and Bouchard C. (2002) Genetics and blood pressure response to
exercise, and its interactions with adiposity. Prev Cardiol 5, 138-144.
Seefeldt V., Malina R. M. and Clark M. A. (2002) Factors affecting levels of
physical activity in adults. Sports M ed 32, 143-168.
Tomling G., Gustavsson P. and Hogstedt C. (1994) Mortality and cancer incidence
in Stockholm fire fighters. Am JIn d M ed 25 ,219-228.
Twisk J. W., Kemper H. C. and van Mechelen W. (2002) The relationship between
physical fitness and physical activity during adolescence and cardiovascular
disease risk factors at adult age. The Amsterdam Growth and Health
Longitudinal Study. Int J Sports M ed 23 Suppl 1, S8-14.
Twisk J. W., Van Mechelen W., Kemper H. C. and Post G. B. (1997) The relation
between "long-term exposure" to lifestyle during youth and young adulthood
and risk factors for cardiovascular disease at adult age. J Adolesc Health 20,
309-319.
Twisk J. W., Kemper H. C., Van Mechelen W. and Post G. B. (2001) Clustering of
risk factors for coronary heart disease, the longitudinal relationship with
lifestyle. Ann Epidemiol 1 1 ,157-165.
van Baak M. A. (1998) Exercise and hypertension: facts and uncertainties. Br J
Sports M ed 32, 6-10.
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Verbeke G. and Molenberghs G. (2000) Springer series in statistics: Linear mixed
models for longitudinak data. Springer-Verlag, New York.
Williams P. T. (2001) Physical fitness and activity as separate heart disease risk
factors: a meta-analysis. M ed Sci Sports Exerc 33, 754-761.
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Abstract (if available)

Abstract The objective of this study was to investigate the longitudinal relationships among physical activity, physical fitness and blood pressure, and to examine the intensity and duration effects of physical activity on blood pressure. Subjects were 2940 males from Los Angels County firefighter cohort, an observational longitudinal study started in 1971 with a maximum follow-up period of 20 years. Physical activity, physical fitness, blood pressure, age, BMI and smoking behavior were repeatedly measured during follow-up. The longitudinal analyses indicated that increasing physical activity and fitness were both inversely related to rates of change in blood pressure. The relationships were consistently observed in different followup periods.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Rates of cognitive decline using logitudinal neuropsychological measures in Alzheimer's disease

PDF

Attrition in a longitudinal drug use prevention study

PDF

Comparison of predicting accuracy of neural networks for censored survival data using generalized Receiver Operating Charactaristic (ROC)-C-Index method

PDF

Associations between physical activities with bone mineral density in postmenopausal women

PDF

Relationship of blood pressure and antihypertensive medications to cognitive change in the BVAIT, WISH, and ELITE clinical trials

PDF

Cross-sectional association of blood pressure, antihypertensive medications, MRI volumetric measures and cognitive function scores in an aging population

PDF

Endometrial cancer in Asian migrants to the United States and their descendants

PDF

Obesity paradox in acute heart failure decompensation

PDF

Air pollution and breast cancer survival in California teachers: using address histories and individual-level data

PDF

Association between latchkey status and smoking behavior in middle school children

PDF

The effects of hormone therapy on carotid artery intima-media thickness and serum lipids by ApoE4 genotype

PDF

Descriptive epidemiology of thyroid cancer in Los Angeles County, 1972-1995

PDF

Associations between isoflavone soy protein (ISP) supplementation and breast cancer in postmenopausal women in the Women’s Isoflavone Soy Health (WISH) clinical trial

PDF

Association between informed decision-making and mental health-related quality of life in long term prostate cancer survivors

PDF

Effects of post-menopausal hormone therapy on arterial stiffness in the ELITE trial

PDF

Body size and the risk of prostate cancer in the multiethnic cohort

PDF

Disparities in colorectal cancer survival among Latinos in California

PDF

Association of oophorectomy with progression of carotid atherosclerosis

PDF

Association of maternal and environmental factors with infant feeding behaviors in a birth cohort study

PDF

Association of subclinical atherosclerosis with plasma B-vitamin, cysteine, homocysteine, and cysteinyl glycine in a cardiovascular disease-free population

Asset Metadata

Creator Liu, Ruolan (author)

Core Title Longitudinal changes in physical activity and physical fitness: associations with blood pressure

School Graduate School

Degree Master of Science

Degree Program Applied biometry and epidemiology

Publication Date 12/09/2020

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Mack, Wendy (committee chair), Azen, Stanley (committee member), Hodis, Howard (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-411049

Unique identifier UC11666550

Identifier 1417929.pdf (filename),usctheses-c89-411049 (legacy record id)

Legacy Identifier 1417929.pdf

Dmrecord 411049

Document Type Thesis

Rights Liu, Ruolan

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA