European Heart Journal (2001) 22, 573–579
doi.10.1053/euhj.2000.2402, available online at http://www.idealibrary.com on
Cardiovascular risk factors and 10-year all-cause
mortality in elderly European male populations
The FINE study
A. Menotti1, I. Mulder1, A. Nissinen2, E. Feskens1, S. Giampaoli3, M. Tervahauta2
and D. Kromhout1
1
Division of Public Health Research, National Institute of Public Health and the Environment, Bilthoven,
The Netherlands; 2Department of Public Health and General Practice, University of Kuopio, Kuopio, Finland;
3
Laboratorio di Epidemiologia e Biostatistica, Istituto Superiore di Sanita’, Rome, Italy
Background This study aims to examine cardiovascular
risk factors in relation to all-cause mortality in elderly
populations of different European countries.
Methods Men aged 65–84 years from defined administrative areas were enrolled in Finland (rural areas of east and
west Finland; n=716), in the Netherlands (the town
of Zutphen; n=887), and in Italy (the rural areas of
Crevalcore and Montegiorgio; n=682). Ten-year all-cause
mortality was studied in relation to measurements taken at
entry: age, systolic blood pressure, HDL- and non-HDLcholesterol, body mass index, heart rate and smoking
habits. Univariate and multivariate analyses were performed with all-cause mortality as the end-point.
Results Ten-year death rates from all causes were higher
in Finland (574 per 1000), lower in the Netherlands (475
per 1000), and Italy (466 per 1000). Age, heart rate
Introduction
The relationship of cardiovascular risk factors to allcause mortality can be easily analysed in population
studies originally designed for research into cardiovascular disease epidemiology. This problem is especially of
interest when the study population has grown old. If one
studies an elderly person who has died the underlying
cause of death may be hard to establish because of the
frequent occurrence of co-morbid conditions.
Within the Seven Countries Study on Cardiovascular
Diseases[1–3], five cohorts (two in Finland, two in Italy
and one in the Netherlands) joined a new study on the
Revision submitted 25 July 2000, and accepted 26 July 2000.
Correspondence: Alessandro Menotti, MD, PhD, Cardioricerca,
Via Adda 87, Rome, Italy 00198.
0195-668X/01/070573+07 $35.00/0
and smoking in all three countries were independently
associated with 10-year all-cause mortality. Non-HDLcholesterol was not related with all-cause mortality. The
observed associations between HDL-cholesterol, systolic
blood pressure, body mass index and all-cause mortality
were dependent on the in- or exclusions of early death.
Conclusion In these elderly men only age, smoking habits
and heart rate were consistently associated with all-cause
mortality.
(Eur Heart J 2001; 22: 573–579, doi:10.1053/euhj.2000.2402)
2001 The European Society of Cardiology
Key Words: Elderly, epidemiology, risk factors, all-cause
mortality, smoking, heart rate.
See page 528 for the Editorial comment on this article
elderly, which started in the mid 1980s, coinciding with
their 25-year re-examination. The men were aged 65 to
84 years and the study was called FINE (Finland, Italy,
Netherlands, Elderly). The adopted baseline field examination of this new study was followed by collection of
mortality data during 10 years of follow-up. A report on
the first 5-year experience[4] suggested that age and
smoking habits were directly associated with future
deaths, while other factors (blood pressure, HDL- and
non-HDL-cholesterol and body mass index) did so
in a parabolic fashion, although many uncertainties
remained about their precise role.
The purpose of the present study is to investigate the
associations between cardiovascular risk factors and
10-year all-cause mortality. We hypothesized that the
measured cardiovascular risk factors are still predictive
of all-cause mortality in these elderly populations.
2001 The European Society of Cardiology
574
A. Menotti et al.
Material and Methods
The FINE study extends the Seven countries Study on
Cardiovascular Diseases beyond the 25 years of followup. It includes the two Finnish cohorts (both rural)
known as East Finland (EF) and West Finland (WF),
the Dutch cohort from the small town of Zutphen (ZU),
and the two rural cohorts located in the villages of
Crevalcore (CR) and Montegiorgio (MO), in Northern
and Central Italy, respectively.
The start of the study was fixed at the time of
the 25-year anniversary re-examination, held in 1984
(Finland) and in 1985 (the Netherlands and Italy), when
the men were aged 65 to 84 years. On that occasion new
rosters were defined by identifying those individuals still
alive in Finland and Italy. In the Dutch area of Zutphen
a new 2/3 statistical sample of men in the same age range
(65 to 84 years) was enrolled and examined for the first
time in 1985 and added to the survivors of the original
Zutphen cohort. This latter cohort dating from 1960 was
a 4/9 statistical sample of the resident male population
aged 40–59 years. Altogether, at the baseline of the
FINE study, 716 men were examined and enrolled in
Finland, 887 in the Netherlands and 682 in Italy, and in
total numbered 2285 men. The participation rate was
92% in Finland, 74% in the Netherlands and 76% in
Italy.
The entry examination of the FINE Study included
measurement of the following risk factors:
Age was measured in years, rounded-off to the nearest
birthday.
Blood pressure was measured in the supine position, at
the end of a physical examination or after 10 min rest,
by a mercury sphygmomanometer (random-zero in the
Netherlands), following the procedure suggested in the
WHO Cardiovascular survey Methods Manual[5].
Total and HDL-serum cholesterol were measured in
blood samples (non-fasting in Italy and Zutphen), using
enzymatic methods. The laboratories were under quality
control of the WHO Lipid Reference Center of Prague
and the CDC Center in Atlanta, and technical details are
reported elsewhere[6]; for the purpose of this analysis, the
variable non-HDL-cholesterol has been created as the
difference between total and HDL-cholesterol, a proxy
for LDL cholesterol.
Height (cm) and weight (kg) were measured in light
undergarments following the procedure suggested by the
WHO Cardiovascular Survey Methods Manual[5]; and
the body mass index (kg . m 2) was calculated. In
Finland height was not measured in 1984; the measurements taken in 1959 at the beginning of the Seven
Countries Study were used instead.
Smoking habits were elicited by a questionnaire originally designed for the Seven Countries Study. The
subjects were classified as never having smoked, exsmokers or current smokers.
Heart rate was measured on a resting ECG record
(average of rate in leads I and V6 in Italy and in the
Netherlands, while in Finland it was measured by pulse
rate count).
Eur Heart J, Vol. 22, issue 7, April 2001
Mortality data were collected during 10 years of
follow-up. Only six participants were lost to follow-up
(two in Italy, four in the Netherlands). Death certificates
were coded using the 9th Revision of the WHO-ICD[7],
by a single investigator. Only all-cause mortality is used
in this paper.
Univariate analyses were performed by computing
age-adjusted death rates in tertile classes or other arbitrary classes of risk factors. Tests of proportions were
calculated for extreme rates for each risk factor. Multivariate analyses were carried out using the Cox proportional hazards model[8] for the prediction of 10-year
all-cause mortality, as a function of risk factors. These
analyses were done for each country and pooled for
all groups together. Dummy variables identified the
cohorts or the countries when more than one cohort
or country was included in a model. The models included age, systolic blood pressure, HDL-cholesterol,
non-HDL-cholesterol, body mass index, smoking
habits, and heart rate. For some variables, both linear
and quadratic terms were added to the models because
the univariate analysis suggested the possibility of
U-shaped relationships.
Results
Mean levels of cardiovascular risk factors at the baseline
of the FINE study are reported in Table 1. Age was
similar in the three countries, but significantly higher
in Italy. Systolic and diastolic blood pressure were
definitely higher in Italy than in Finland and the
Netherlands. Mean levels of total cholesterol were the
same in Finland and the Netherlands, but significantly
lower in Italy. The same was true for non-HDLcholesterol. Levels of HDL-cholesterol were significantly
higher in Italy, and lower in the other two countries.
Body mass index was somewhat higher in Italy, intermediate in Finland and lowest in the Dutch men. Mean
heart rate was similar in Finland and in Italy and
significantly higher in the Netherlands. Prevalence of
smokers was highest in the Netherlands, intermediate in
Italy and definitely lowest in Finland. The reverse was
true for ex-smokers. Ten-year all-cause death rate was
highest in Finland (574 per 1000), and similar in the
Netherlands (475 per 1000) and Italy (466 per 1000).
All-cause age-adjusted death rates were computed in
tertile classes of risk factors and in three arbitrary classes
for smoking habit. Relative risks were computed comparing rates in tertile 1 with those in tertiles 2 and 3
(Table 2). Systolic and diastolic blood pressure were not
related to all-cause mortality in any of the three
countries. In Finland, total and non-HDL-cholesterol
were inversely related to all-cause mortality. A similar
but not statistically significant association was observed
for HDL cholesterol. Total HDL- and non-HDLcholesterol were not associated with all-cause mortality
in Italy and the Netherlands. Body mass index was
inversely related to all-cause mortality in all three
Risk factors and all-cause mortality in elderly men
575
Table 1 Mean level and standard deviation (SD) of risk factors at baseline.
Standard errors are given for smoking habits
Risk factor
Age (years)
Systolic BP (mmHg)
Diastolic BP (mmHg)
Total chol. (mmol . l 1)
HDL-chol. (mmol . l 1)
Non-HDL-chol. (mmol . l 1)
Heart rate (beats . min 1)
Body mass index (kg m 2)
Never smokers (%)
Ex smokers (%)
Current smokers (%)
Finland
Netherlands
Italy
Mean
SD
Mean
SD
Mean
SD
71·6
154·0
86·7
6·1
1·2
4·9
67·9
25·7
26·5
55·3
18·3
5·1
22·7*
11·2*
1·3
0·3
1·3
12·1*
4·1
1·6*
1·9
1·4*
71·5
151·1
85·4
6·1
1·1
5·0
75·6
25·5
18·4
51·4
30·2
5·3†
21·5†
11·5†
1·1†
0·3†
1·1†
14·1†
3·2†
1·30†
1·7†
1·5
72·0
166·7
91·3
5·9
1·3
4·6
68·6
25·9
29·5
43·9
26·6
4·5‡
22·4‡
11·1‡
1·1‡
0·3‡
1·2‡
13·5
3·6
1·75
1·9‡
1·7‡
*P<0·05 Finland vs Netherlands.
†P<0·05 Netherlands vs Italy.
‡P<0·05 Italy vs Finland.
Table 2 Relative risk of 10-year age adjusted mortality from all causes as a function
of risk factors measured at entry
Risk factor
Systolic blood pressure
Diastolic blood pressure
Total cholesterol
HDL-cholesterol
Non-HDL-cholesterol
Body mass index
Heart rate
Smoking habits
Tertiles or
classes
Finland
Netherlands
Italy
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
Never
Ex
Current
1·00
0·91
0·99
1·00
0·93
0·95
1·00
0·81
0·81*
1·00
0·90
0·86
1·00
0·87
0·82*
1·00
0·85
0·80*
1·00
1·01
1·10
1·00
1·10
1·28†
1·00
1·05
1·11
1·00
0·91
1·02
1·00
0·96
0·95
1·00
0·95
0·95
1·00
0·89
0·94
1·00
0·91
0·93
1·00
1·15
1·33*
1·00
0·99
1·32†
1·00
0·97
1·12
1·00
0·97
1·15
1·00
1·00
0·95
1·00
0·83
0·88
1·00
1·18
0·98
1·00
0·86
0·91
1·00
1·20
1·34*
1·00
1·13
1·21
*P<0·05 between tertile 1 and tertile 3.
†P<0·05 between never smokers and current smokers.
countries. A significant difference between rates in tertile
3 vs tertile 1 was only found in Finland. The highest
death rates were found in tertile 3 of heart rate in the
three countries, and the difference with tertile 1 was
significant in the Netherlands and Italy. In all three
countries current smokers had the highest death rate.
The results of multivariate analyses are presented
in Table 3. In Finland smoking habits and heart rate
were positively associated with all-cause mortality and
HDL-cholesterol and body mass index inversely. In the
Netherlands age, heart rate and smoking habits were
also positively related to all-cause mortality and in Italy
only age and heart rate.
In the pooled analyses for the three countries age,
heart rate and smoking were positively associated
with 10-year all-cause mortality (Table 4). HDLcholesterol and body mass index were inversely associated. Systolic blood pressure was marginally associated
Eur Heart J, Vol. 22, issue 7, April 2001
576
A. Menotti et al.
Table 3 Solutions of the proportional hazards model predicting all-cause 10-year mortality as a function of risk factors
measured at entry examination
Country
Finland
Denominator=670; cases=369
Netherlands
Denominator=879; cases=415
Italy
Denominator=640; cases=273
Risk factor
Coefficient
S.E.
P
Coefficient
S.E.
P
Coefficient
S.E.
P
Age
Systolic BP
HDL chol.
Non-HDL-chol.
Heart rate
BMI
Smokers
West Finland
Montegiorgio
0·1014
0·00008
0·6095
0·0420
0·0108
0·04235
0·4412
0·2077
—
0·0109
0·0025
0·1786
0·0459
0·0042
0·0147
0·1592
0·1090
—
0·0001
0·9716
0·0006
0·3636
0·0099
0·0041
0·0056
0·0568
—
0·1053
0·0033
0·1637
0·0038
0·0094
0·0357
0·3918
—
—
0·0096
0·0023
0·1880
0·0498
0·0035
0·0169
0·1452
—
—
0·0001
0·1490
0·3839
0·9388
0·0070
0·0350
0·0070
—
—
0·1318
0·0051
0·1690
0·0098
0·0183
0·0205
0·2179
—
0·0460
0·0140
0·0028
0·1917
0·0564
0·0042
0·0180
0·1660
—
0·1230
0·0001
0·0682
0·3779
0·8626
0·0001
0·2513
0·1894
—
0·7083
BP=blood pressure; BMI=body mass index.
Table 4 Pooled analyses of the association between risk factors and 10-year all-cause mortality in three countries
(denominator=2189; cases=1057)
Risk factor
Age
Systolic BP
HDL chol.
Non-HDL-chol.
Heart rate
BMI
Smokers
West Finland
Zutphen
Crevalcore
Montegiorgio
Coefficient
S.E.
P value
HR
0·1098
0·0025
0·3275
0·0180
0·0122
0·0320
0·3774
0·2649
0·5237
0·6345
0·5978
0·0064
0·0014
0·1070
0·0290
0·0023
0·0094
0·0892
0·1062
0·0953
0·1243
0·1154
0·0001
0·0797
0·0022
0·5337
0·0001
0·0006
0·0001
0·0126
0·0001
0·0001
0·0001
1·73
1·02
0·94
0·99
1·06
0·97
1·45
0·77
0·59
0·53
0·55
95% CI
1·63,
0·99,
0·90,
0·96,
1·04,
0·95,
1·22,
0·62,
0·49,
0·41,
0·44,
1·84
1·05
0·98
1·02
1·08
0·99
1·73
0·94
0·71
0·68
0·69
Differences for
HR estimates
5 years
10 mmHg
0·2 mmol . l 1
0·5 mmol . l 1
5 beats . min 1
1 unit
Yes–No
vs East Finland
vs East Finland
vs East Finland
vs East Finland
BP=blood pressure; HR=hazard ratio; BMI=body mass index.
and non-HDL-cholesterol was not related with all-cause
mortality. Men in West Finland were at lower risk
compared with men in East Finland. The lowest risk was
observed in the Dutch and Italian cohorts.
Besides linear models parabolic relationships were
also studied by adding quadratic terms for the risk
factors, which suggested a non-linear relationship in
univariate analyses e.g. HDL and non-HDL-cholesterol.
Consistent associations for the linear and quadratic
component of the associations between HDL and nonHDL-cholesterol and 10-year all-cause mortality were
found in Finland, the Netherlands and the pool of the
three countries (results not shown).
In order to study the influence of morbidity on the
association between cardiovascular risk factors and
10-year all-cause mortality we excluded deaths occurring
in the first 5 years of follow-up (Table 5). In this analysis
age, systolic blood pressure, heart rate and cigarette
smoking were positively associated with all-cause mortality. Non-HDL-cholesterol and body mass index were
not related.
Another side-analysis was conducted exploring the
possible role of drug therapy, that is common in these
Eur Heart J, Vol. 22, issue 7, April 2001
older men, on coefficients of cardiovascular risk factors.
Only antihypertensive drugs were systematically and
uniformly coded in the three national groups and therefore only these were considered. Prevalence of antihypertensive treatment was common overall in Finland
and Italy (34% of all men) but relatively rare in the
Netherlands (12%). All-cause death rates in 10 years
were systematically higher among those taking antihypertensive drugs compared with those who did not
(67% vs 51% in Finland; 58% vs 46% in the Netherlands;
50% vs 41% in Italy). The inclusion of antihypertensive
treatment as a dummy variable in proportional hazards
equations produced positive and highly significant coefficients for this covariate in Finland, the Netherlands
and in the pool, but not in Italy. The hazards ratios
ranged between 1·21 in Italy and 1·83 in Finland.
However, no substantial influence was noted on the
magnitude of risk factor coefficients that maintained
the same significant levels as in the equation without the
inclusion of antihypertensive treatment. A further analysis segregating those who took drugs from those who did
not, did not reach any clear-cut conclusion due to the
small numbers involved.
Risk factors and all-cause mortality in elderly men
577
Table 5 All-cause mortality between 5 and 10 years of follow-up as a function of risk factors measured at entry, in the
pool of the three countries with exclusion of first 5-year deaths (denominator 1674; cases 543)
Risk factor
Age
Systolic BP
HDL chol.
Non-HDL-chol.
Heart rate
BMI
Smokers
West Finland
Zutphen
Crevalcore
Montegiorgio
Coefficient
S.E.
HR
0·1123
0·0035
0·2137
0·0649
0·0097
0·0182
0·5160
0·1998
0·3881
0·5321
0·5758
0·0092
0·0020
0·1494
0·0414
0·0033
0·0133
0·1245
0·1540
0·1394
0·1717
0·1631
1·75
1·04
0·96
0·97
1·05
0·98
1·67
0·82
0·68
0·59
0·56
95% CI
1·60,
1·00,
0·90,
0·93,
1·02,
0·96,
1·31,
0·61,
0·52,
0·42,
0·41,
1·92
1·18
1·02
1·01
1·18
1·01
2·14
1·11
0·89
0·82
0·77
Differences for
HR estimates
5 years
10 mmHg
0·2 mmol . l 1
0·5 mmol . l 1
5 beats . min 1
1 unit
Yes–No
vs East Finland
vs East Finland
vs East Finland
vs East Finland
BP=blood pressure; HR=hazard ratio; BMI=body mass index.
Discussion
The results of the present study suggest that age, heart
rate and smoking are important predictors of all-cause
mortality in elderly men. HDL-cholesterol and body
mass index were inversely related to all-cause mortality,
but these associations were no longer statistically significant when the men who died during the first 5 years of
follow-up were excluded. In contrast, systolic blood
pressure was only positively associated with all-cause
mortality after exclusion of early death. Non-HDLcholesterol was unrelated to all-cause mortality.
The overall findings of the present study differ from
those found after the first 5 years of follow-up[4]. On that
occasion most risk factors other than age, heart rate and
cigarette smoking had a distinct parabolic relationship
with all-cause mortality. In this long-term experience,
non-HDL-cholesterol was not predictive of all-cause
mortality. HDL and body mass index were inversely
related to all-cause mortality, but these associations
were no longer statistically significant when the first
5 years of death were excluded. This suggests that with
increasing age at death an early parabolic relationship of
some risk factors with all-cause mortality may become
inverse and even non-existent after excluding early
death. These results emphasize the great importance of
sufficient follow-up periods and the exclusion of morbid
conditions by excluding early death in prospective
studies on risk factors and mortality in the elderly.
Cigarette smoking was consistently associated with
all-cause mortality after 5 and 10 years of follow-up in
elderly men from Finland, the Netherlands and Italy.
These results are consistent with those found in a 5-year
follow-up study carried out in the U.S.A. among persons
aged 65 and older[9]. In an Italian study on persons aged
80 and over no association was found between cigarette
smoking and all-cause mortality[10]. The results of the
present study show that in elderly men the risk by
former smokers did not differ significantly from that of
those who had never smoked. This suggests that even in
elderly men there is a health benefit in stopping smoking.
The coronary risk of former smokers declines quickly
after stopping, although a slight excess may persist for
many years[11].
Heart rate is an important risk factor for cardiovascular mortality[12]. The present study shows that
heart rate is an important independent predictor for
both 5- and 10-year all-cause mortality in elderly men
from Finland, the Netherlands and Italy. A recent paper
from the Framingham Heart Study showed that reduced
heart rate variability in the elderly is an independent risk
factor for all-cause mortality[13]. There is increasing
evidence that heart rate (variability) is an independent
risk factor for all-cause mortality, indicating a protective
effect of physical activity. A low heart rate is associated
with a high level of physical activity[14].
In the present study, non-HDL-cholesterol, an indicator of LDL-cholesterol, was not related to all-cause
mortality in any analysis. A direct relationship between
total cholesterol and all-cause mortality was only found
in cases with a very long follow-up period between
cholesterol measurement and events or when measurements were taken at relatively young ages[15,16]. In the
case of measurements at old age no, uncertain or inverse
associations were found[10,17–22]. The direction of the
association may depend upon age, gender and ethnicity[23]. In this study, parabolic relationships were shown
between both HDL- and non-HDL-cholesterol on one
side and all-cause mortality on the other, but they were
lost when early deaths were excluded.
HDL-cholesterol was inversely related to all-cause
mortality in Finland and in the pooled data of Finland,
the Netherlands and Italy. However, this association
was influenced by morbid conditions present in these
elderly men at the baseline study. In an analysis in which
early death was excluded, the inverse relationship
between HDL-cholesterol and all-cause mortality was
no longer statistically significant in healthy elderly men.
In the present study, systolic blood pressure was
not associated with all-cause mortality in elderly men
from Finland, the Netherlands and Italy. However,
in a pooled analyses excluding early death, a positive
Eur Heart J, Vol. 22, issue 7, April 2001
578
A. Menotti et al.
significant association between systolic blood pressure
and all-cause mortality was observed. These results
indicate that in healthy elderly men systolic blood
pressure may still be a predictor of all-cause mortality.
Also in some other reports a positive relationship was
observed between blood pressure and all-cause mortality[24,25]. However, no relationships, conflicting results
and inverse associations have also been reported[26–30].
This may be due to differences in age, health status
and treatment in different populations of elderly
persons[31,32].
Our data suggested an inverse relationship between
body mass index and all-cause mortality. This inverse
relationship was no longer statistically significant after
exclusion of early death. In the NHANES follow-up
study, a curvilinear association was found between
body weight and mortality in older persons[33]. The
Copenhagen Study showed that the inverse relationship
between body weight and mortality was partly explained
by recent changes in weight or co-morbidity[34]. In fact,
low body weight in the elderly may be the consequence
of a serious disease, psychiatric disorders followed by
anorexia, social factors, changes in food regulation, taste
and olfactory sensitivity, loss of teeth, drug-nutrient
interactions, etc.[35]. The results of the present study
suggest that in healthy elderly men, body mass index is
probably not related to all-cause mortality.
The results of the present study showed that the health
status of the elderly is important in judging relationships
between cardiovascular risk factors and mortality. Comorbidity occurs frequently in the elderly and is critical
in the understanding of the impact of risk factors on
longevity[36]. In the present analyses the problem has
been tackled by excluding early death in order to select a
relatively healthy subgroup of the population. More
detailed analyses on co-morbidity have been submitted
elsewhere. At the same time, after this introductory
exploration of the role of traditional cardiovascular risk
factors on all-cause mortality, attention is moving toward the association between the same factors and
functional capacity, quality of life, disability, selfsufficiency and depression, since information on these
topics is available in this study.
The relatively common use of antihypertensive drugs
did not show any real influence on the magnitude of risk
factor coefficients. The paradox of the positive and
significant coefficients estimated for drug use is in contrast with the outcome of many clinical trials[37], but is
easily explained by the selection made by the practising
physician in deciding when and to whom to prescribe
antihypertensive drugs.
In conclusion age, heart rate and cigarette smoking
were important predictors of all-cause mortality in the
following 10 years in these European cohorts of elderly
men. Non-HDL-cholesterol was not associated with
all-cause mortality. Unequivocal associations were not
found for HDL-cholesterol, blood pressure and body
mass index. The observed associations for these risk
factors were dependent on in- or excluding early death.
The results of the present study suggest that lifestyleEur Heart J, Vol. 22, issue 7, April 2001
related risk factors e.g. smoking and heart rate (an
indicator for physical activity) remain important predictors of longevity in elderly men in different European
cultures.
References
[1] Keys A, ed. Coronary heart disease in seven countries.
Circulation 1970; 41 (Suppl 1): 1–211.
[2] Keys A, ed. Seven Countries: a multivariate analysis of death
and coronary heart disease. Cambridge (MA): Harvard Univ
Press, 1980: 1–381.
[3] Menotti A, Keys A, Kromhout D et al. Inter-cohort differences in coronary heart disease mortality in the 25-year
follow-up of the Seven Countries Study. Eur J Epidemiol
1993; 9: 527–36.
[4] Menotti A, Kromhout D, Nissinen A et al. Short-term allcause mortality and its determinants in elderly males populations in Finland, the Netherlands and Italy: the FINE study.
Prev Med 1996; 25: 319–26.
[5] Rose G, Blackburn H. Cardiovascular survey methods.
WHO, Geneva, 1968.
[6] Kromhout D, Nissinen A, Menotti A, Bloemberg B,
Pekkanen J, Giampaoli S. Total and HDL cholesterol and
their correlates in elderly men in Finland, Italy and the
Netherlands. Am J Epidemiol 1990; 131: 855–63.
[7] WHO. International classification of diseases. 9th Revision.
Geneva: WHO, 1975.
[8] Cox DR. Regression models and life tables. J R Stat Soc 1972;
B43: 187–220.
[9] LaCroix AZ, Lang J, Scherr P et al. Smoking and mortality
among older men and women in three communities. N Engl J
Med 1991; 324: 1619–25.
[10] Casiglia E, Spolaore P, Ginocchio G et al. Predictors of
mortality in very old subjects aged 80 years or over. Eur J
Epidemiol 1993; 9: 577–86.
[11] Omenn GS, Anderson KW, Kronmal RA, Vlietstra RE. The
temporal pattern of reduction of mortality risk after smoking
cessation. Am J Prev Med 1990; 6: 251–7.
[12] Platini P. Heart rate as a cardiovascular risk factor. Eur Heart
J Suppl 1999; 1 (Suppl B): B3–B9.
[13] Tsuji H, Venditti FJ Jr, Manders ES et al. Reduced heart
rate variability and mortality risk in an elderly cohort. The
Framingham Heart Study. Circulation 1994; 90: 878–83.
[14] Bijnen FCH, Feskens EJM, Caspersen CJ et al. Physical
activity and cardiovascular risk factors among elderly men in
Finland, Italy and the Netherlands. Am J Epidemiol 1996;
143: 553–61.
[15] Shipley MJ, Pocock SJ, Marmot MG. Does plasma cholesterol concentration predict mortality from coronary heart
disease in elderly people? 18 year follow up in Whitehall study.
BMJ 1991; 303: 89–92.
[16] Smith GD, Shipley MJ, Marmot MG, Rose G. Plasma
cholesterol concentration and mortality. The Whitehall Study.
JAMA 1992; 267: 70–6.
[17] Woo J, Chan SM, Mak YT. Swaminathan R. Biochemical
predictors of short term mortality in elderly residents of
chronic care institutions. J Clin Pathol 1989; 42: 1241–5.
[18] Ives DG, Bonino P, Traven ND, Kuller LH. Morbidity and
mortality in rural community-dwelling elderly with low total
serum cholesterol. J Gerontol 1993; 48: M103–7.
[19] Kronmal RA, Cain KC, Ye-Z, Omenn GS. Total serum
cholesterol levels and mortality risk as a function of age. A
report based on the Framingham data. Arch Intern Med 1993;
153: 1065–73.
[20] Krumholz HM, Seeman TE, Merrill SS et al. Lack of association between cholesterol and coronary heart disease mortality and morbidity and all-cause mortality in persons older
than 70 years. JAMA 1994; 272: 1335–40.
Risk factors and all-cause mortality in elderly men
[21] Weverling-Rijnsburger AWE, Blauw GJ, Lagaay AM, Knook
DL, Meinders AE, Westendorp RGJ. Total cholesterol and
risk of mortality in the oldest old. Lancet 1997; 350: 1119–23.
[22] Higgins M, Keller JB. Cholesterol, coronary heart disease,
and total mortality in middle-aged and elderly men and
women in Tecumseh. Ann Epidemiol 1992; 2: 69–76.
[23] White AD, Hames CG, Tyroler HA. Serum cholesterol and
20-year mortality in black and white men and women aged 65
and older in the Evans County Heart Study. Ann Epidemiol
1992; 2: 85–91.
[24] Glynn RJ, Field TS, Rosner B, Hebert PR, Taylor JO,
Hennekens CH. Evidence for a positive linear relation between blood pressure and mortality in elderly people. Lancet
1995; 345(8953): 825–9.
[25] Vokonas PS, Kannel WB, Cupples LA. Epidemiology and
risk of hypertension in the elderly. The Framingham Study.
J Hypertens 1988; 6 (Suppl (1)): S3–S9.
[26] Langer RD, Ganiats TG, Barrett-Connor EL. Factors associated with paradoxical survival at higher blood pressures in the
very old. Am J Epidemiol 1991; 134: 29–38.
[27] Langer RD, Criqui MH, Barrett-Connors EL, Klauber MR,
Ganiats TG. Blood pressure change and survival after age 75.
Hypertension 1993; 22: 551–9.
[28] Ekbom T, Londholm L, Oden A et al. Blood pressure does
not predict mortality in the elderly. J Hypertension 1988; 6
(Suppl 4): S626–8.
[29] Busby WJ, Campbell AJ, Robertson MC. Is low blood
pressure in elderly people just a consequence of heart disease
and frailty? Age Ageing 1994; 23: 69–74.
579
[30] Boshuizen HC, Izaks GJ, Van Buuren S, Lighart GJ.
Blood pressure and mortality in people aged 85 and over:
Community based study. Br Med J 1998; 316: 1780–4.
[31] Smith WM. The case for treating hypertension in the elderly.
Am J Hypertens 1988; 1: 173S–8S.
[32] Glynn RJ, Field TS, Satterfield S et al. Modification of
increasing systolic blood pressure in the elderly during the
1980s. Am J Epidemiol 1993; 138: 365–79.
[33] Cornoni-Huntley JC, Harris TB, Everett DF et al. An overview of body weight of older persons, including the impact on
mortality. The National Health and Nutrition Examination
Survey 1 — Epidemiologic Follow-up Study. J Clin Epidemiol
1991; 44: 743–53.
[34] Suadicani P, Hein HO, Gyntelberg F. Weight changes and
risk of ischaemic heart disease for middle aged and elderly
men. An 8-year follow-up in the Copenhagen Male study.
J Cardiovasc Risk 1997; 4: 25–32.
[35] Fischer J, Johnson MA. Low body weight and weight loss in
the aged. J Am Diet Assoc 1990; 90: 1697–706.
[36] Cornoni-Huntley JC, Foley DJ, Guralnik JM. Co-morbidity
analysis: a strategy for understanding mortality, disability and
use of health care facilities of older people. Int J Epidemiol
1991; 20 (Suppl 1): S8–17.
[37] Pitt B, Julian DG, Pocock SJ, eds. Clinical trials in cardiology.
London, WB Saunders Co Ltd, 1997.
Eur Heart J, Vol. 22, issue 7, April 2001