406
Effects of Traditional Coronary Risk Factors
Rates of Incident Coronary Events in a
Low-Risk Population
on
The Adventist Health Study
Gary E. Fraser, MB, ChB, PhD, MPH, FRACP; T. Martin Strahan, MBBS, DrPH;
J. Sabate, MD, DrPH; W. Lawrence Beeson, MSPH; and D. Kissinger, PhD
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
Background. California Seventh-Day Adventists have lower mortality rates from coronary heart disease
(CHD) than other Californians. Associations between traditional risk factor and CHD events have not
been reported previously for Adventists.
Methods and Results. A cohort study allowed 6 years of follow-up of 27,658 male and female California
Seventh-Day Adventists. Data collected included age, sex, physician-diagnosed hypertension and diabetes
mellitus, body height, weight, previous and current cigarette smoking habits, and current exercise habits.
Incident cases of definite myocardial infarction (MI) and definite fatal CHD were diagnosed according to
recognized criteria. Both stratified and proportional hazards analyses demonstrated that in this low-risk
population, the above traditional coronary risk factors exhibit their usual associations with risk of CHD
events. It was noted that exercise had a strong negative association with fatal CHD events (relative risks
[RR], 1.0, 0.66, and 0.50 with increasing exercise) but no association with risk of MI (either nonfatal or
all cases). Conversely, obesity was much more clearly associated with MI (RR, 1.0, 1.18, and 1.83 with
increasing tertiles of obesity) than with fatal events. The importance of the risk factors was similar in both
sexes, except that the effect of cigarette smoking seemed more pronounced in women.
Conclusions. The epidemiology of coronary heart disease in this low-risk California population appears
to be at least qualitatively similar to that seen in other groups. There was evidence that the effects of
exercise and obesity may differ depending on whether fatal CHD and MI (either all MI or nonfatal alone)
is the end point. (Circulation 1992;86:406-413)
KEY WoRDs * hypertension * smoking * obesity * exercise * populations * Adventist Health
Study
It has been well established from previous studies
that California Seventh-Day Adventists are at low
risk for fatal coronary heart disease (CHD) events
compared with other Californians.' Reasons that have
been postulated include the nonsmoking status of Adventists, the sizable proportion of members who adhere
to a lacto-ovo vegetarian diet, the possibility of increased psychosocial support, and the higher-than-average educational status. Of interest is the observation
that even when compared with the nonsmoking members of a contemporary California study population
having a slightly higher socioeconomic status, the Adventists showed substantially lower age-specific mortality rates from coronary disease in both men and
women.2
From the Center for Health Research, Loma Linda University
School of Public Health, Loma Linda, Calif.
Supported by National Institutes of Health grant 5-RO1-HL26210. Work done by J.S. occurred during the tenure of a research
fellowship from the American Heart Association, California
Affiliate.
Address for correspondence: Gary E. Fraser, Center for Health
Research, Room 2008, Nichol Hall, Loma Linda University, Loma
Linda, CA 92350.
Received February 5, 1991; revision April 6, 1992.
Adventists are an epidemiologically attractive study
population because they encompass a wide range of
dietary habits but are characterized by very little current use of tobacco or consumption of alcohol. The
former increases statistical power to investigate dietary
hypotheses, and the latter reduces the possibilities of
confounding by these factors. Also, their interest in
health should result in members having greater awareness of their health habits and willingness to complete
long and tedious questionnaires. Thus, more complete
and accurate exposure data may be anticipated. The
Adventist Health Study (AHS) is a cohort study of
31,208 California Seventh-Day Adventists who were
followed to detect fatal and nonfatal coronary and
cancer events between 1977 and 1982.3 An important
premise underlying the design of this study was that the
results associating various exposures to risk of CHD in
this special population should be applicable to the
general population. We postulate that although there
undoubtedly are complex socioreligious selective influences influencing membership in the Seventh-Day Adventist church, these influences would be unlikely to
affect the biological and metabolic mechanisms of exposure-disease associations. Evidence to support this
contention may be provided by examining the effect of
Fraser et al Risk Factors and Coronary Events in Adventists
variables usually found to be associated with risk of
CHD in other populations.
Thus, we investigated the effect of smoking habits,
hypertension, obesity, physical activity, and a diagnosis
of diabetes mellitus on the risk of both definite myocardial infarction and definite fatal CHD events in men and
women of this study population.
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
Methods
The AHS population and design have been described
in detail elsewhere.3 Briefly, in 1974, a census questionnaire was mailed to all Adventist households in the state
of California. This census identified 59,081 Adventists
from different ethnic backgrounds who were 25 years
old or older. In August 1976, a detailed lifestyle questionnaire was mailed to all cohort members. The response rate in non-Hispanic whites was 75%. Information obtained relevant to CHD included a personal
history of previous physician-diagnosed heart disease,
diabetes, or hypertension; self-reported height and
weight; previous and current cigarette smoking; and
exercise and dietary habits. Exercise was categorized as
low, moderate, or high; this represents a cross-classification of two questions relating to occupational and
leisure activities. The following questions were asked:
"Outside of your usual work or daily activities, do you
usually get at least 15 minutes of vigorous exercise three
or more times per week?" (There followed a list of
possible vigorous activities for the subject to check if he
or she undertook these at least three times each week
for 15 minutes on each occasion.) "Does your usual
daily work or responsibilities involve vigorous activities
similar to those listed in the previous question?" (Responses were "very often," "frequently," "occasionally," "rarely," or "never.") "High" exercise reflects a
high occupational and/or high leisure activity status;
"moderate" exercise includes moderate leisure activity
and low or moderate occupational activity or low leisure
activity but moderate occupational activity; and "low"
exercise includes subjects with both low leisure and
occupational activities.
For a period of 6 years (1977 through 1982), annual
questionnaires were mailed to all participants as a
screening mechanism to identify new cases of CHD.
Information on any hospitalization within the previous
12-month period was requested by this questionnaire,
and permission to review any relevant medical records
was obtained. Valid returns from these questionnaires
were obtained from an average of 97% of study subjects
each year and from 99.5% in the final year, when any
missing data from previous years also were collected.
Pertinent portions of the hospital records were microfilmed by study field representatives to allow confirmation of the diagnosis by AHS physicians. All ECGs were
microfilmed, and cardiac enzyme results were abstracted to a special form. Follow-up in this fashion was
completed for 97% of the participants. The ECGs were
Minnesota coded4 by both a Minnesota Coding Laboratory-trained technician and a physician. Any discrepancies between the two coders were resolved by consensus. Periodically, sample batches of ECGs were sent to
the Minnesota Coding Laboratory as an aid to quality
control. Medical records were reviewed without knowledge of any exposure variable.
407
A diagnosis of nonfatal myocardial infarction was
confirmed if the international diagnostic criteria5 were
met. In summary, these require a typical ECG change
or elevation of cardiac enzyme levels plus either prolonged cardiac pain or static ECG abnormalities. Fatal
CHD also was defined by international diagnostic criteria5 as either "definite fatal myocardial infarction" or
"other definite fatal CHD." Definite fatal myocardial
infarction required death within 30 days of a myocardial
infarction confirmed by hospital records as described
above or fresh myocardial infarction recorded at
autopsy.
Other definite fatal CHD required International Classification of Disease, 9th Revision code 410-414 as underlying or immediate cause of death on the death
certificate provided there were no other likely lethal
causes on the certificate. In addition, it was required
that there was a previous history of CHD, autopsy
findings of severe coronary disease, or symptoms compatible with a coronary cause of death. In no case was
the cause on a death certificate accepted without supplementary evidence. Deaths were determined for the
total study population by computer-assisted linkage
with the California death certificate files,6,7 the National
Death Index, and church records. For any possible case
of fatal CHD, relatives were contacted to elicit information pertinent to the circumstances of death. Also,
when available, autopsy reports were procured.
The few subjects who experienced a new definite
nonfatal myocardial infarction and a subsequent definite fatal CHD event could contribute an end point to
analyses of both syndromes. In no case was the same
individual counted twice in the same analysis. Although
such a fatal event was not strictly incident, the exposures
were documented before the first event and could not
have been influenced by the subject's knowledge of his
or her CHD. The 179 subjects with possible coronary
events were included as noncases in the analysis. Analyses excluding them did not change the results in any
important way.
We confine this report to data from non-Hispanic
white Adventists. A total of 31,208 non-Hispanic white
Adventists returned the questionnaire in 1976. All
subjects who at study baseline had a previous or unknown history of heart disease are excluded. This left
27,658 subjects for these analyses.
The effect of the main exposures of interest were first
assessed using the person-years version of the stratified
Mantel-Haenszel procedure.8 Because our study population was large, it was possible to stratify on five or six
variables each at two or three levels. In addition,
multivariate proportional hazards analyses are reported
with all variables of interest included in the model. A
particular variable often was represented by one or two
dummy terms in the model indicating different levels of
exposure. Statistical significance was assessed by a likelihood ratio test9 comparing log-likelihood functions
between models with and without the term(s) representing the risk factor of interest. The adequacy of the
proportional hazards assumption was checked by the
inspection of log-log plots.
On average, approximately 1-2% of data were missing for the variables of this analysis. For the stratified
analyses, all subjects with missing data on any of the
relevant variables were excluded. For the Cox propor-
408
Circulation Vol 86, No 2 August 1992
TABLE 1. Selected Characteristics of the Adventist Health Study
Population in 1976
Men
Women
(n = 17,282)
(n = 10,376)
51.6±16.0
53.5±16.6
Age (years, mean±SEM)
3.6
4.7
Diabetes (%)
15.8
Hypertension (%)
22.6
68.7
Never smoker (%)
85.7
29.1
12.9
Past smoker (%)
2.2
Current smoker (%)
1.4
24.9±3.5
24.3±4.7
Quetelet index (±SEM)*
31.6
46.8
Low exercise (%)
22.3
17.1
Moderate exercise (%)
46.1
36.2
High exercise (%)
*Quetelet index equals weight (kg)/(height [m])2.
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
tional hazards analyses, the method described by Woodward et al10 was used to conserve subjects for analysis.
For all variables except Quetelet index, there was no
indication of important bias with this method. Thus,
persons with missing data for Quetelet index were
excluded from these analyses to avoid biased estimates
of effect.
Results
The characteristics of the population are shown in
Table 1. The population has a wide age range (25-99
years) with a substantial number of elderly persons
(26.4% of persons were >65 years old). The proportion
of current smokers is very small, as expected, but a
sizable minority have been past smokers. During the six
years of follow-up, 203 cases of incident definite myocardial infarction and 302 incident definite fatal CHD
events were documented. There is some overlap; 48
cases of definite myocardial infarction subsequently
contributed to definite fatal CHD events. The age and
sex distributions and incidence rates of these events are
shown in Table 2. It is noteworthy that although there
are more fatal events than myocardial infarctions, if
attention is restricted to persons less than 80 years of
age, there are modestly more cases of myocardial infarction (148) than deaths (126).
Results of Mantel-Haenszel analyses, stratified on
age and traditional risk factors (i.e., diabetes mellitus,
hypertension, cigarette smoking, physical activity, and
Quetelet index of obesity) aside from the exposure of
interest are shown in Table 3 for the end point of
definite myocardial infarction. For both sexes, the effects of diabetes mellitus, hypertension, and passive
cigarette smoking are similar to those reported by
numerous other cohort studies, although our ability to
investigate current smoking is very limited as only one
male case and one female case were observed in this
predominantly nonsmoking population. An independent effect of obesity is seen and appears to be somewhat greater in men than women. There is no significant
trend in risk of definite myocardial infarction with level
of physical activity for either sex.
Similar analyses are presented in Table 4 for definite
fatal CHD. Again, strong associations with diabetes
mellitus, hypertension, and past cigarette smoking are
shown. The effects of current cigarette smoking are
particularly strong in the women, although for men only
one fatal case is observed. Therefore, the confidence
interval is very wide. However, the results for physical
activity and obesity contrast with those for definite
myocardial infarction. For fatal CHD, increasing physical activity is associated with a marked decrease in risk
in both men and women, whereas obesity does not show
a clear independent association with risk for either sex.
We also explored the association of the Quetelet
index of obesity with disease end points when adjusted
for age and sex alone. For definite myocardial infarction, the relative risks from lowest to highest tertiles are
1.00, 2.07, and 3.14 (p<0.0001), and for fatal CHD the
corresponding figures are 1.00, 1.03, and 1.28 (p=NS).
It often is instructive to compare results from alternative methods of analysis. Thus, Tables 5 and 6 present
analyses of the same data using the Cox proportional
hazards method. As can be seen, the point estimates of
the corresponding relative risks and hazard ratios are
very similar, and the confidence intervals overlap substantially. That a few minor differences do occur is to be
expected and is discussed below briefly.
Discussion
and
consistent associations between hypertenStrong
sion, cigarette smoking, and diabetes mellitus and risk
of ischemic heart disease events among California Seventh-Day Adventists are similar to those reported from
a wide variety of other populations."1-14 This study
includes nearly 200,000 person-years of observation and
is one of the few cohort studies in which data were
collected from large numbers of both men and women
(62.5% of subjects were female), thus allowing an
equitable comparison between the sexes. Excellent validity has been documented for selected exposure variables when compared with face-to-face interviews.3
However, a potential weakness of the data is that all
information was gathered by self-reporting. Thus, the
variables hypertension and diabetes were self-reported,
although ostensibly physician diagnosed. We note that
TABLE 2. Cases of Incident Coronary Heart Disease 1977-1982: The Adventist Health Study
Definite fatal coronary heart disease
Definite myocardial infarction
Age
(years)
No.
0
18
60
58
Men
Incidence*
No.
2
5
41
118
25-44
...
45-64
7.42
65-79
48.09
80+
154.12
*Per 10,000 person-years of follow-up.
Women
Incidence*
0.70
1.33
16.81
140.89
No.
1
38
58
25
Men
Incidence*
0.54
15.67
46.49
66.44
No.
1
9
41
30
Women
Incidence*
0.35
2.40
16.81
35.82
Fraser et al Risk Factors and Coronary Events in Adventists
409
TABLE 3. Mantel-Haenszel Stratified Analyses Associating Risk of Definite Myocardial Infarction With Traditional
Coronary Risk Factors: The Adventist Health Study
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
Men
Women
Total
Variable
Cases/PY
RR (95% CI)
Cases/PY
RR (95% CI)
(RR [95% CI])
Diabetes
Absent
88/51,803
1.00*
51/82,184
1.00t
1.00*
Present
11/1,825
2.08 (1.10-3.93)
12/3,842
2.34 (1.22-4.49)
2.21 (1.40-3.48)
Hypertension
Absent
69/45,466
1.00*
23/67,153
1.00T
1.00L
Present
30/8,161
1.57 (1.01-2.46)
40/18,873
2.38 (1.37-4.11)
1.90 (1.35-2.66)
Cigarette smoking
Never
1.00
55/37,192
51/73,804
1.00
1.00*
Past
1.42 (0.95-2.13)
43/15,362
11/11,042
1.90 (0.97-3.72)
1.52 (1.07-2.16)
Current
0.94 (0.13-6.90)
1/1,074
1/1,181
2.43 (0.26-22.67)
1.36 (0.32-5.76)
Physical activity
Low
1.00
24/17,007
31/40,128
1.00
1.00
Medium
1.71 (1.00-2.94)
35/11,951
11/14,633
0.87 (0.43-1.75)
1.33 (0.88-2.00)
0.89 (0.53-1.50)
High
40/24,669
21/31,266
0.97 (0.55-1.71)
0.93 (0.63-1.36)
Quetelet index
Low
1.OO
10/13,634
1.00*
15/36,739
1.00*
Medium
2.64 (1.32-5.28)
1.17 (0.57-2.42)
46/21,723
15/25,681
1.89 (1.16-3.06)
2.79 (1.35-5.77)
High
43/18,270
2.08 (1.08-3.98)
33/23,608
2.40 (1.48-3.89)
PY, Person-years of follow-up; CI, confidence intervals; RR, relative risk.
Stratified by age, sex (where appropriate), and the other risk factors included in this table aside from the exposure of
interest. *p.0.05, tp<0.01, *p<0.001, testing the null hypothesis of equal effects in all exposure categories.
adequate validity has been documented when a similar
approach was applied to the diagnosis of hypertension
in well-educated populations.15-'7 Our California Adventist study population not only has a great interest in
their own health but also is well educated (72% of men
and 62% of women have some college education). For
diabetes mellitus, it is the usual experience that a
proportion of prevalent cases have not been previously
diagnosed.18 This would predict that the methods of this
study may have missed a number of cases of diabetes.
However, despite this, a relatively high prevalence of
diabetes was indicated at baseline (see Table 1) in a
TABLE 4. Mantel-Haenszel Analyses Associating Risk of Definite Fatal Coronary Heart Disease With Traditional Coronary Risk Factors: The Adventist Health Study
Men
Women
Total
RR
Variable
RR (95% CI)
(RR [95% CI])
Cases/PY
(95% CI)
Cases/PY
Diabetes
1.00
Absent
1.00t
l.00t
92/51,803
119/82,184
1.57 (0.85-2.89)
Present
1.82 (1.12-2.97)
1.17 (1.72-2.52)
13/1,825
20/3,842
Hypertension
Absent
Present
Cigarette smoking
Never
Past
Current
1.00*
63/45,466
42/8,161
70/67,153
69/18,873
1.00*
1.46 (1.03-2.05)
1.00*
2.26 (1.49-3.43)
55/37,192
49/15,362
1/1,074
1.00
1.44 (0.97-2.15)
1.41 (0.19-10.62)
120/73,804
14/11,042
5/1,181
1.00*
1.00*
1.67 (0.94-3.00)
7.44 (2.62-21.16)
1.51 (1.08-2.10)
4.35 (1.77-10.67)
1.72 (1.31-2.25)
Physical activity
1.00
1.00*
Low
1.00*
37/17,007
92/40,128
0.70 (0.42-1.18)
0.61 (0.37-0.98)
0.65 (0.45-0.92)
Medium
26/11,951
21/14,633
0.61 (0.39-0.96)
0.41 (0.26-0.64)
0.49 (0.36-0.67)
High
42/24,669
26/31,266
Quetelet index
1.00
1.00
1.00
Low
23/13,634
63/36,739
1.31 (0.77-2.23)
0.85 (0.56-1.28)
1.00 (0.73-1.38)
Medium
41/21,723
38/25,681
1.46 (0.84-2.53)
0.79 (0.52-1.22)
1.01 (0.73-1.40)
High
41/18,270
38/23,608
PY, Person-years of follow-up; CI, confidence intervals; RR, relative risk.
Stratified by age, sex (where appropriate), and the other risk factors included in this table aside from the exposure of
interest. *p<0.05, tp.0.01, *p<0.001, testing the null hypothesis of equal effects in all exposure categories.
410
Circulation Vol 86, No 2 August 1992
TABLE 5. Proportional Hazards Analysis Associating Traditional Coronary Risk Factors With Risk of Definite
Myocardial Infarction: The Adventist Health Study
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
Variable
Sex (M, 1; F, 0)
Age (years)
25-44
45-64
65-79
80+
Diabetes
Hypertension
Cigarette smoking
Never
Past
Current
Physical activity
Low
Medium
High
Quetelet index
Low
Medium
High
CI, confidence intervals.
Men
(person-years, 55,007;
hazard ratio [95% CI])
Women
(person-years, 89,346;
hazard ratio [95% CI])
0.23 (0.10-0.52)
0.23 (0.05-1.00)
1.00*
4.79 (2.64-8.70)
7.65 (3.31-17.70)
2.15* (1.33-3.47)
1.00*
2.01 (1.29-3.14)
4.38 (2.34-8.22)
2.11* (1.12-4.00)
1.69* (1.09-2.62)
Total
(person-years, 144,353;
hazard ratio [95% CI])
2.87* (2.11-3.91)
0.23 (0.12-0.48)
1.00*
2.84 (2.02-4.00)
5.24 (3.25-8.45)
2.96* (1.78-4.93)
2.15* (1.37-3.37)
2.19* (1.59-3.02)
1.00
1.51 (1.01-2.25)
0.96 (0.13-6.94)
1.00
1.77 (0.93-3.34)
2.50 (0.34-18.28)
1.00
1.51 (1.07-2.12)
1.34 (0.33-5.44)
1.00
1.50 (0.90-2.50)
0.90 (0.55-1.49)
1.00
0.84 (0.44-1.63)
0.82 (0.49-1.40)
1.00
1.17 (0.79-1.72)
0.81 (0.57-1.17)
1.00
1.64 (0.79-3.43)
2.02 (0.98-4.16)
1.00t
0.74 (0.35-1.57)
1.71 (0.91-3.20)
l.00t
1.18 (0.72-1.95)
1.83 (1.14-2.93)
*p<0.05, tpO.Ol, *p<O.00l, likelihood ratio test of null hypothesis that all exposure categories have equal effects.
population documented to have a low mortality ascribed to diabetes.19 This may indicate a fairly complete
ascertainment. A similar self-reported diagnostic approach for diabetes has been used by others2021 in
similar studies. In most epidemiological research, it is of
greatest importance that good specificity of diagnosis be
obtained. The cases not diagnosed will have been
considered nondiabetics, but they represent so numerically small a group that they hardly perturb the population of true nondiabetics.
Some differences between the proportional hazards
analyses and the stratified analyses are to be expected.
Proportional hazards analyses take account of the time
during follow-up that person-years are contributed and
assume that the hazard ratio remains constant during
this period, whereas these are not considered in the
Mantel-Haenszel approach. In addition, we note that
current smokers compose only 2% of our cohort. Therefore, it is probable that small numbers have resulted in
biased or unstable maximum likelihood estimates for
this variable in the proportional hazards model. The
experience of persons aged 25-44 years with only two
cases of myocardial infarction and two cases of definite
fatal CHD contribute very little to the Mantel-Haenszel
analyses but contribute with the other noncases to the
likelihood of the proportional hazards analysis. Finally,
the two methods handled missing data in different ways
(see "Methods"), resulting in a minor difference (23%) in person-years between the two types of analyses.
Data from certain other low-risk populations had
indicated that the male-to-female ratio for CHD mortality was much less than for higher-risk populations.22
We did not find this below the age of 80 years in the
Adventists. Between the ages of 45 and 79 years, the
male-to-female incidence ratio varied between 2.77 and
6.53 for the different specific age, sex, and end point
groups. Qualitatively, the strength of association between the various risk factors and risk of disease seemed
quite similar for the two sexes. Current smoking in
women was associated with a striking elevation in risk of
fatal CHD, but evaluation of the effect in men was
hindered by the fact that only one current smoker
experienced a fatal CHD event. The effect of past
smoking on both end points also seemed somewhat
stronger for women.
Past smokers had elevated point estimates of risk
varying from 1.42 to 1.90 for the different sexes and
CHD end points. Although we did not record the time
since quitting smoking, the average time since becoming
a Seventh-Day Adventist in past smokers was 23.7 years.
Because most Adventists quit smoking at the time of
admission to the church, the average time since quitting
would be at least this long. Despite this relatively long
average period since quitting, the effect of past smoking
could still be detected on risk of CHD events.
There were apparently quite different associations of
physical activity with definite myocardial infarction and
definite fatal CHD. Although there appeared to be a
marked and significant independent protective effect on
fatal CHD for both sexes, no consistent trends were
found for risk of incident definite acute myocardial
infarction. A formal comparison of the difference in the
exercise association between fatal and nonfatal events
was borderline significant (p<0.07). A review of observational studies reveals that although a few investigators
have found significant negative associations between
Fraser et al Risk Factors and Coronary Events in Adventists
411
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
TABLE 6. Proportional Hazards Analysis Associating Traditional Coronary Risk Factors With Risk of Definite
Fatal Coronary Heart Disease: The Adventist Health Study
Total
Women
Men
(person-years, 144,353;
(person-years, 89,346;
(person-years, 55,007;
hazard ratio [95% CI])
hazard ratio [95% CI])
hazard ratio [95% CI])
Variable
1.67* (1.26-2.21)
Sex (M, 1; F, 0)
Age (years)
0.13 (0.04-0.44)
0.28 (0.06-1.31)
0.07 (0.01-0.34)
25-44
1.00*
1.00*
1.00*
45-64
9.17 (6.02-13.95)
7.71 (4.58-12.97)
13.95 (6.63-29.33)
65-79
16.71 (9.23-30.26)
79.68 (37.86-167.67)
36.20 (23.53-55.69)
80+
1.86* (1.30-2.67)
1.71* (1.06-2.76)
2.05t (1.18-3.55)
Diabetes
1.78* (1.37-2.30)
1.50* (1.07-2.09)
2.22* (1.49-3.29)
Hypertension
Cigarette smoking
1.00t
1.00*
Never
1.00
1.57 (1.15-2.14)
1.82 (1.05-3.17)
1.48 (1.01-2.16)
Past
3.12 (1.37-7.10)
0.71 (0.10-5.16)
8.05 (3.26-19.90)
Current
Physical activity
1.00*
Low
1.00
1.00*
0.66 (0.48-0.92)
Medium
0.75 (0.46-1.22)
0.64 (0.40-1.01)
0.50 (0.37-0.68)
0.60 (0.39-0.93)
0.46 (0.30-0.70)
High
index
Quetelet
1.00
1.00
Low
1.00
0.77 (0.44-1.36)
0.74 (0.49-1.11)
0.74 (0.54-1.02)
Medium
1.11 (0.65-1.90)
0.89 (0.60-1.32)
0.98 (0.72-1.33)
High
CI, confidence intervals.
*p<0.05, tp<O.Ol, $p<0.001, likelihood ratio test of null hypothesis that all exposure categories have equal effects.
physical activity and acute myocardial infarction, the
magnitude of the effect23-26 usually has been relatively
small and less than that for fatal CHD27-32 or for the
combination of fatal and nonfatal events.33,34 It is of
interest that the same appears to be true in randomized
trials of cardiac rehabilitation. The two largest North
American trials found no evidence of benefit on risk of
myocardial infarction.35,36 Only the latter study reported
on fatal CHD events and found suggestive evidence of
benefit here. A meta-analysis of nearly 20 randomized
studies of cardiac rehabilitation also found suggestive
evidence of benefit for fatal events but no such evidence
of benefit for nonfatal myocardial infarction.37
Atherosclerosis contributes in a major way to both
nonfatal myocardial infarction and fatal coronary
events, and the effects of physical activity on risk factors
such as blood lipids, blood pressure, obesity, and fibrinolysis are well described. Reasons for a possible stronger association with fatal CHD are unclear, although it
is not surprising that there should be etiologic differences between these two syndromes as the pathophysiology is not identical. A substantial proportion of CHD
deaths are primarily arrhythmic,38 whereas arrhythmia
is not commonly involved in the etiology of nonfatal
myocardial infarction. It is of interest that both human39A40 and animal41 data indicate that reduced vagal
activity, a risk factor for malignant ventricular arrhythmias,42 can be normalized by regular physical activity.
With our data, we were concerned that physical
inactivity, measured before any diagnosis of CHD, may
have represented a preclinical manifestation of the
disease rather than a preceding cause. Such an effect
would be expected to weaken as follow-up proceeded,
with the later events being relatively far removed from
the baseline characteristics. Cox proportional hazards
analyses similar to those of Table 6 but separate for
fatal CHD occurring in 1977-1979 and 1980-1982 (both
sexes combined) showed relative risks (proceeding from
the lowest to the highest exercise categories) of 1.00,
0.77, and 0.59 and of 1.00, 0.71, and 0.58, respectively,
for the two follow-up periods. Thus, no weakening of
effect was noted, which further enforces that the exercise habits preceded the clinical disease.
Diabetes mellitus has been associated with increased
risk of CHD in many epidemiological studies. An interesting feature is that others often have found effects to
be higher in women than in men, with the female
diabetics losing much of their sex advantage in risk for
CHD.4344 However, this did not appear to be true of a
population of Pima Indians who were at low risk for
CHD but at high risk for diabetes.45 Some have found
that diabetic women are more likely to die or develop
congestive heart failure after an acute myocardial infarction than were diabetic men.46 The pathophysiology
underlying these sex differences is unclear and is not
obviously explained by differences in traditional risk
factors between diabetic men and women.47 It is of
interest that in the Adventist population, the relative
risks of diabetics for both definite myocardial infarction
and definite fatal CHD are similar in men and women.
We were unable to adjust for lipid differences, but this
has not explained the sex differences in other studies.37
Our study classified only known diabetics but does not
differ in this respect from other epidemiological studies
also showing the sex difference.2' Thus, in this low-risk
population, the advantage of being female is not lost
even in diabetics.
412
Circulation Vol 86, No 2 August 1992
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
In our data, the Quetelet index of obesity was associated with both definite myocardial infarction and
definite fatal CHD when adjusted only for age and sex.
The association appeared monotonic for myocardial
infarction but U shaped for fatal events. This latter
association persisted weakly in the Cox multivariate
analyses but was not seen in stratified analyses, and no
results were statistically significant. Proportional hazards models that included terms for Quetelet index of
obesity and its square in addition to other covariates
were evaluated. In neither men nor women nor for
either end point was the squared term statistically
significant. A formal comparison of the difference in the
obesity associations between fatal and nonfatal events
was borderline significant (p<0.06).
Most cohort studies of CHD have shown age- and
sex-adjusted associations between some measure of
obesity with risk of CHD events,48-50 but this usually has
not persisted after adjustment for other traditional risk
factors.464851 The implication is that either there is
confounding between obesity and other traditional risk
factors or some of the other risk factors are intervening
variables between obesity and disease risk, as is known
to be so. A few studies have documented significant
independent associations between obesity and CHD
syndromes,52-55 although this sometimes occurred after
very long follow-up or with adjustment for only a subset
of the traditional risk factors. Recent interest in the
effect of distribution of fat shows that this has predictive
ability beyond that of "weight for height" measures.
This variable was not measured in the present study.
Data on associations between traditional risk factors
and incidence of CHD in low-risk populations are
uncommon. One such population is the Japanese living
in Japan. A cohort study has associated serum cholesterol and blood pressure but not smoking with CHD
incidence at both 6- and 16-year follow-ups.56 A casecontrol study of myocardial infarction in Japan57 also
significantly associated serum cholesterol, previous history of hypertension, and cigarette smoking with risk.
However, obesity was not clearly related to risk. Studies
of Pima Indians find a low risk of CHD but diabetes still
to be a risk factor as all events occurred in diabetics.58
The number of events is too small to clearly assess the
effect of other traditional risk factors. The well-known
Seven Countries Study has presented data for 10-year
follow-up on relatively low-risk Southern European
cohorts in Italy, Greece, and Yugoslavia.59 Although
some analyses suffered from small numbers, there was
good evidence that high blood pressure and high serum
cholesterol were associated with increased risk. Physical
inactivity predicted increased risk of CHD death but
showed a weaker nonsignificant association when all
CHD events are combined. Cigarette smoking may have
been hazardous, but this approached statistical significance only for CHD death. No clear evidence was found
associating obesity with risk of CHD.
In summary, there is general agreement between data
from other low-risk populations and our findings in
California Adventists that the traditional risk factors are
associated with CHD risk in low-risk populations. This
provides added evidence that findings in the California
Adventist population regarding less-well-established
risk factors such as foods in the diet or psychosocial
factors will also have application to other populations.
Acknowledgments
We acknowledge the excellent advice received from Drs.
Ralph Paffenbarger, Ronald Prineas, and Stephen Fortmann,
who served as consultants to this study.
References
1. Lemon FR, Walden RT: Death from respiratory system disease
among Seventh-day Adventist men. JAMA 1966;198:117-126
2. Phillips RL, Kuzma JW, Beeson WL, Lotz T: Influence of selection
versus lifestyle on risk of fatal cancer and cardiovascular disease
among Seventh-day Adventists. Am J Epidemiol 1980;112:296-314
3. Beeson WL, Mills PK, Phillips RL, Andress M, Fraser GE:
Chronic disease among Seventh-day Adventists, a low-risk group:
Rationale, methodology and description of the population. Cancer
1989;64:57-81
4. Prineas RJ, Crow RS, Blackburn H: The Minnesota Code Manual of
Electrocardiographic Findings. London, John Wright, PSG Inc,
1982
5. Gillum RF, Fortman SP, Prineas RJ, Kottke TE: International
diagnostic criteria for acute myocardial infarction and stroke. Am
Heart J 1984;108:150-158
6. Kuzma JW, Beeson WL: The relationship of lifestyle characteristics to mortality among California Seventh-day Adventists. Proceedings of the 19th National Meeting of the Public Health Conference
on Records and Statistics. DHHS Publication No. (PHS) 81-1214,
pp 195-200, Hyattsville, Md, National Center for Health Statistics,
US Dept HHS, Dec 1980
7. Beeson WL, Fraser GE, Mills PK: Validation of record linkage to
2 California population-based tumor registries in a cohort study.
Proceedings of the 1989 Public Health Conference on Records and
Statistics. DHHS Publication No. (PHS) 90-1214, 1989,
pp 196-201, Hyattsville, Md, National Center for Health Statistics,
US Dept HHS, Nov 1989
8. Rothman KJ: Modern Epidemiology. Boston, Little, Brown & Co,
1986, pp 196, 202
9. Wilks SS: The large sample distribution of the likelihood ratio for
testing composite hypotheses. Ann Math Stat 1938;9:60
10. Woodward M, Smith WCS, Tunstall-Pedoe H: Bias from missing
values: Sex differences in implication of failed venepuncture for
the Scottish Heart Health Study. Int J Epidemiol 1991;20:379-383
11. Wilhelmsen L, Wedel H, Tibblin G: Multivariate analysis of risk
factors for coronary heart disease. Circulation 1973;48:950-958
12. Gordon T, Sorlie P, Kannel WB: The Framingham Study: An
Epidemiologic Investigation of Cardiovascular Disease. Section 27.
Coronary Heart Disease, Atherothrombotic Brain Infarction, Intermittent Claudication. A Multivariate Analysis of Some Factors
Related to their Incidence. Washington, DC, US Government Printing Office, Publication No. 1740-0320, 1971
13. Kagan A, Gordon T, Rhoads GG, Schiffman JC: Some factors
related to coronary heart disease in Honolulu Japanese men: The
Honolulu Heart Study. Int J Epidemiol 1975;4:271-279
14. Rosenberg L, Miller DR, Kaufman DW, Helmrich SP, Van de
Carr S, Stolley PD, Shapiro S: Myocardial infarction in women
under 50 years of age. JAMA 1983;250:2801-2806
15. Paffenbarger RS, Wing AL, Hyde RT, Jung DL: Physical activity
and incidence of hypertension in college alumni. Am J Epidemiol
1983;117:245-257
16. Paffenbarger RS, Thorne CM, Wing AL: Chronic disease in former
college students: VIII. Characteristics in youth predisposing to
hypertension in later years. Am J Epidemiol 1968;88:25-32
17. Fiebach N, Hebert PR, Stampfer MJ, Colditz GA, Willett WC,
Rosner B, Speizer FE, Hennekens CH: A prospective study of high
blood pressure and cardiovascular disease in women. Am J Epide-
miol 1989;130:646-654
18. Dawber TR: The Framingham Study. Cambridge, Mass, Harvard
University Press, 1980, p 192
19. Snowdon DA, Phillips RL: Does a vegetarian diet reduce the
occurrence of diabetes? Am J Public Health 1985;75:507-512
20. Abbott RD, Donahue RP, MacMahan SW, Reed DM, Yano K:
Diabetes and the risk of stroke: The Honolulu Heart Program.
JAMA 1987;257:949-952
21. Schumacher MC, Hunt SC, Williams RR: Interactions between
diabetes and family history of coronary heart disease among adults
with diabetes in Utah. Epidemiology 1990;1:298-304
22. Acheson RM: The etiology of coronary heart disease. Yale J Biol
Med 1962;35:143-170
Fraser et al Risk Factors and Coronary Events in Adventists
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
23. Morris JN, Pollard R, Everitt MG, Chave SPW: Vigorous exercise
in leisure-time: Protection against coronary heart disease. Lancet
1980;2:1207-1210
24. Paffenbarger RS, Wing AL, Hyde RT: Physical activity as an index
of heart attack risk in college alumni. JEpidemiol 1978;108:161-175
25. Salonen JT, Puska P, Tuomilehto J: Physical activity and risk of
myocardial infarction, cerebral stroke and death. Am J Epidemiol
1982;115:526-537
26. Holme I, Helgeland A, Hjermann I, Leren P, Lund-Larsen P: Four
and two-thirds years incidence of coronary heart disease in middleaged men: The Oslo Study. Am J Epidemiol 1980;112:149-159
27. Slattery ML, Jacobs DR, Nichaman MZ: Leisure time physical
activity and coronary heart disease death: The US Railroad Study.
Circulation 1989;79:304-311
28. Salonen JT, Slater JS, Tuomilehto J, Rauramaa R: Leisure time
and occupational physical activity: Risk of death from ischemic
heart disease. Am J Epidemiol 1988;127:87-94
29. Holme I, Helgeland A, Hjermann I, Leren P, Lund-Larsen PG:
Physical activity at work and at leisure in relation to coronary risk
factors and social class. Acta Med Scand 1981;209:277-283
30. Brand RJ, Paffenbarger RS, Sholtz RI, Kampert JB: Work activity
and fatal heart attack studied by multiple logistic risk analysis. Am
J Epidemiol 1979;110:52-62
31. Kannel WB: Habitual level of physical activity and risk of coronary
heart disease: The Framingham Study. Can Med Assoc J 1967;96:
811-812
32. Pekkanen J, Nissinen A, Marti B, Tuomilehto J, Punsar S, Karvonen MJ: Reduction of premature mortality by high physical
activity: A 20-year follow-up of middle-aged Finnish men. Lancet
1987;1:1473-1477
33. Garcia-Palmieri MR, Cruz-Vidal M, Havlik RJ: Increased physical
activity: A protective factor against heart attacks in Puerto Rico.
Am J Cardiol 1982;50:749-755
34. Kannel WB, Sorlie P: Some health benefits of physical activity.
Arch Intern Med 1979;139:857-861
35. Rechnitzer PA, Cunningham DA, Andrew GM, Buck CW, Jones
NL, Kavanagh T, Oldridge NB, Parker JO, Shephard RJ, Sutton
JR, Donner AP: Relation of exercise to the recurrence rate of
myocardial infarction in men: Ontario Exercise-Heart Collaborative Study. Am J Cardiol 1983;51:65-74
36. Shaw LW: Effects of a prescribed supervised exercise program on
mortality and cardiovascular morbidity in patients after a myocardial infarction. Am J Cardiol 1981;48:39
37. O'Connor GT, Buring JE, Usuff S, Yusuf S, Goldhaber SZ, Olmstead EM, Paffenbarger RS, Hennekens CH: An overview of randomized trials of rehabilitation with exercise after myocardial
infarction. Circulation 1989;80:234-244
38. Goldstein S, Landis JR, Leighton R, Ritter G, Vasu CM, Lantis A,
Serokman R: Characteristics of the resuscitated out-of-hospital
cardiac arrest victim with coronary heart disease. Circulation 1981;
64:977-984
39. Goldsmith R, Steinman RC, Apfelbaum MA, Bigger JT: Endurance training enhances parasympathetic activity: Possible explanation for the favorable effects of exercise on the survival of postinfarction patients. (abstract) JAm Coll Cardiol 1992;19:371A
40. Goldberger AL, Mietus JE, Rigney DR, Wood ML, Forney SM:
Short-term bedrest deconditioning causes a loss of complex heart
rate variability. (abstract) JAm Coll Cardiol 1992;19:371A
41. Billman GE, Schwartz PJ, Stone HL: The effects of daily exercise
on susceptibility to sudden cardiac death. Circulation 1984;69:
1182-1189
413
42. Schwart PJ, LaRovere MT, Vanoli E: Autonomic nervous system
and sudden cardiac death. Circulation 1992;85(suppl I):I-77-1-91
43. Kannel WB, McGee DL: Diabetes and cardiovascular disease.
JAMA 1979;241:2035-2038
44. Barrett-Connor E, Wingard DL: Sex differential in ischemic heart
disease mortality in diabetics: A prospective population-based
study. Am J Epidemiol 1983;118:489-496
45. Nelson RG, Sievers ML, Knowler WC, Swinburn BA, Pettit DJ,
Saad MF, Liebow IM, Howard BV, Bennett PH: Low incidence of
fatal coronary heart disease in Pima Indians despite high prevalence of non-insulin-dependent diabetes. Circulation 1990;81:
987-995
46. Savage MP, Krotewski AS, Kenien GG, Lebeis MP, Christlieb R,
Lewis SM: Acute myocardial infarction in diabetes mellitus: Significance of congestive heart failure as a prognostic factor. Am J
Cardiol 1988;62:665-669
47. Heyden S, Heiss G, Bartel AG, Hames CG: Sex differences in
coronary mortality among diabetics in Evans County, Georgia.
J Chron Dis 1980;33:265-273
48. Keys A, Aravanis C, Blackburn H, Van Buchem FSP, Buzina R,
Djordjevic BS, Fidanza F, Karvonen MJ, Menotti A, Puddu V,
Taylor HL: Coronary heart disease: Overweight and obesity as risk
factors. Ann Intern Med 1972;77:15-27
49. Lapidus L, Bengtsson C, Larsson B, Pennert K, Rybo E, Sjostrom
L: Distribution of adipose tissue and risk of cardiovascular disease
and death: A 12 year follow up of participants in the population
study of women in Gothenburg, Sweden. Br Med J 1984;289:
1257-1261
50. Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR:
Diabetes, blood lipids, and the role of obesity in coronary heart
disease risk for women. Ann Intern Med 1977;87:393-397
51. Rosenman RH, Brand RJ, Sholtz RI, Friedman M: Multivariate
prediction of coronary heart disease during 8.5 year follow-up in
the Western Collaborative Group Study. Am J Cardiol 1976;37:
903-909
52. Dyer AR, Stamler J, Berkson DM, Lindberg HA: Relationship of
relative weight and body mass index to 14-year mortality in the
Chicago Peoples Gas Company Study. J Chron Dis 1975;28:109-123
53. Hubert HB, Feinleib M, McNamara PM, Castelli WP: Obesity as
an independent risk factor for cardiovascular disease: A 26-year
follow-up of participants in the Framingham Heart Study. Circulation 1983;67:968-977
54. Manson JE, Colditz GA, Stampfer MJ, Willet WC, Rosner B,
Monson RR, Speizer FE, Hennekens CH: A prospective study of
obesity and risk of coronary heart disease in women. N Engl J Med
1990;322:882-889
55. Schroll M: A longitudinal epidemiological survey of relative weight
at age 25, 50 and 60 in the Glostrup population of men and women
born in 1914. Dan Med Bull 1981;28:106-116
56. Szatrowski TP, Peterson AV, Shimizu Y, Prentice RL, Mason
MW, Fukunaga Y, Kato H: Serum cholesterol, other risk factors,
and cardiovascular disease in a Japanese cohort. J Chron Dis 1984;
37:569-584
57. Hyams L, Segi M, Archer M: Myocardial infarction in the Japanese: A retrospective study. Am J Cardiol 1967;20:549-554
58. Ingelfinger JA, Bennett PH, Liebow IM, Miller M: Coronary heart
disease in the Pima Indians. Diabetes 1976;25:561-565
59. Keys A: Seven Countries: A Multivariate Analysis of Death and
Coronary Heart Disease. Cambridge, Mass, Harvard University
Press, 1980
Effects of traditional coronary risk factors on rates of incident coronary events in a
low-risk population. The Adventist Health Study.
G E Fraser, T M Strahan, J Sabaté, W L Beeson and D Kissinger
Downloaded from http://circ.ahajournals.org/ by guest on June 18, 2017
Circulation. 1992;86:406-413
doi: 10.1161/01.CIR.86.2.406
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1992 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://circ.ahajournals.org/content/86/2/406
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/