Journul ofCienmlohgy: MEDICAL SCIENCES
1998, Vol. 53A. No. 5. M347-M350
Copyright 1998 by The Gemntological Society of America
Longevity and Gray Hair, Baldness, Facial Wrinkles,
and Arcus Senilis in 13,000 Men and Women:
The Copenhagen City Heart Study
Peter Schnohr, J0rgen Nyboe, Peter Lange, and Gorm Jensen
Epidemiological Research Unit, The National University Hospital, Copenhagen, Denmark.
Background. We have previously reported that men who look older than their contemporaries have a significantly
higher risk for myocardial infarction. The purpose of this study was to investigate whether persons with pronounced
aging signs such as graying of hair, baldness, or facial wrinkles are prone to a shorter life span compared to their
contemporaries.
Methods. In the Copenhagen City Heart Study comprising a random sample of 20,000 men and women, we also
recorded, in addition to cardiovascular risk factors, data on signs of aging: extent of gray hair, baldness, facial wrinkles, and arcus senilis (corneal arcus). During 16 years of follow-up, 3,939 persons (1,656 women and 2,283 men) had
died. The Cox regression model for proportional hazards, which included age as an explanatory variable, was used for
descriptive analysis of the correlation between these aging signs and all-cause mortality.
Results. We found no correlation between the mortality and the extent of graying of the hair, or baldness or facial
wrinkles in either of the sexes, irrespective of age. A single exception was observed in a small subgroup of men with
no gray hair. They had a slightly, but significantly, lower mortality than the rest [relative risk (RR) = .81, 95% confidence interval (CI) .67-.98; p < .05]. The presence of arcus senilis was significantly correlated with a shorter life span
in women (RR = 1.25, 95% CI 1.08-1.46; p < .01). For men the same tendency was found, but the correlation was not
statistically significant.
Conclusion. We conclude that the degrees of graying of the hair, baldness, and facial wrinkles are not predictive of
a shorter life span in men and women in the Copenhagen City Heart Study.
H
AIR graying, baldness, and degree of facial wrinkles
are prominent characteristics of the aging process in
humans. It could be questioned if these signs of aging also
reflect aging processes of the whole organism, in particular
the extent of atherosclerosis. If this is the case, it may be
postulated that elderly looking persons have a shorter life
span than others with the same chronological age.
We have previously reported that the above mentioned
signs of aging were significantly correlated with development of myocardial infarction in men even after statistical
control for the most important established cardiovascular
risk factors (1). In addition, Herrera and colleagues, using
data from the Framingham Study, have reported that the
progression of baldness was significantly associated with
coronary heart disease and all-cause mortality in men (2).
The present study was undertaken to test whether the
degree of gray hair, baldness, facial wrinkles, and arcus
senilis—a line of opacity close to the border of the cornea—
after adjustment for age is predictive of a shorter life span.
Thus, the study is purely descriptive and therefore no
covariates other than age have been included in the analysis.
MATERIALS AND METHODS
Participants.—The Copenhagen City Heart Study is a
prospective cardiovascular population study comprising a
random sample of 20,000 men and women 20 years and
older (3). In 1976-78 the participants were invited to a
health examination at The National University Hospital in
Copenhagen. The overall participation rate was 74%. The
present analysis comprises 13,359 persons (7,270 women
and 6,089 men) 30-79 years old.
The information about deaths was obtained from the
Danish National Population Register, which is almost
100% complete, as all Danish citizens can be followed by
means of a Civil Registration System number.
The number of deaths during the 16 years of follow-up
was 3,939 (1,656 women and 2,283 men). The distribution
of participants and number of deaths by age and sex is
shown in Table 1.
Examination procedures.—Established procedures and
examinations for cardiovascular epidemiology surveys
were employed (4). In addition, we added examination of
four aging variables: gray hair, baldness, facial wrinkles,
and arcus senilis. The assessment of aging variables was
performed in 1976-78 by trained laboratory technicians
who were unaware of the subjects' age and disease profile.
However, we were not able to assess each of the aging variables in a blind fashion. Men and women were examined in
a similar way. Prevalence data of aging variables in our
total study population have been published previously (1,3).
M347
SCHNOHR ETAL.
M348
The extent of gray hair was described as (a) no gray
hairs; (b) few gray hairs; (c) moderate gray hairs; or (d)
completely gray/white hair. We have previously published
findings that, after the age of 50, humans are seldom without gray hair. At this age, around 50% of men and 30% of
women are moderately or totally gray haired. As up to
25% of women dye their hair, it is impossible to investigate differences in the extent of gray hair between the
sexes (1,3). It was not possible for us to evaluate axillary
hair color, as the majority of Copenhagen women remove
underarm hair.
The extent of baldness was described in the frontoparietal region as (a) no bald triangle; (b) bald triangle but
>3 cm in front of ear; or (c) bald triangle but <3 cm in front
of ear. The baldness was also described in the crown-top
Statistics.—The subject of the analysis was the hazard or
instantaneous rate of death. We have used the Cox regression model for proportional hazards (7). The period of observation was counted from the date of the initial examination until the person had survived for 16 years or—if died
before—until the date of death. Separate analyses were performed for men and women.
As the purpose of the analysis was to examine whether
the aging variables are predictors of death, we have analyzed the relation between each variable and the death hazard separately using age as the only covariable (8). To
investigate whether the extent of prediction changes with
age, an interaction term (Variable X Age) was included in a
separate analysis.
region as (a) thick hair; (b) partly thin hair; (c) bald spot, or
RESULTS
(d) bald top and front. This assessment was based on
Hamilton's classification (5) and modified by us (1,3).
As expected we found a marked difference between the
sexes: men lose their hair, but women keep it. At the age of
50 years around 25% of men, but no women, have a bald
top (1,3).
The extent of facial wrinkles in the crow's-foot area, lateral to the canthus of the right eye, was described as (a) no
wrinkles; (b) one to three shallow wrinkles less than 1.5
cm in length; (c) one to six shallow wrinkles less than 3
cm in length; (d) several prominent wrinkles not more than
4 cm in length plus shallow wrinkles; (e) several prominent wrinkles more than 4 cm in length plus shallow wrinkles. This classification is a minor modification of Daniell's (6). The extent of prominent wrinkles increases by
age in both sexes, although at a somewhat faster rate for
men (1,3).
The arcus senilis was described as (a) no arcus, (b) half
arcus, or (c) complete arcus in the right eye. The extent of
arcus increases by age and again at a somewhat faster rate
for men. Approximately 25% of men and 15% of women
50 years of age have a half or complete arcus senilis (3).
Gray hair.—Table 2 shows the association between the
extent of graying of the hair and the mortality expressed as
percentage of deaths during 16 years of follow-up by sex
and age groups.
For women, the Cox analysis showed no significant difference in mortality rates across the different categories of
hair-graying.
For men, there was no difference in mortality between
those with few, moderate, or totally gray hair irrespective
of age, but men with no gray hair had a slightly but significantly lower mortality than the rest. The relative risk was
.81, 95% CI .67-.98; p < .05. Although this trend was statistically significant, its impact on the population was limited as the percentage of men 50 years and older with no
gray hair was small: 50-59 yrs old: 6%, 60-69 yrs old: 2%,
70-79 yrs old: 1% (1,3).
Baldness.—Table 3 and 4 show, respectively, the extent
of frontoparietal and crown-top baldness and the mortality
by sex and age groups in the Copenhagen City Heart Study
population.
For both sexes the Cox regression models showed no
significant correlation between the extent of frontal and
crown-top baldness and mortality.
Table 1. Number of Participants in This Analysis and of Deaths
After 16 Years of Follow-up by Sex and Age
Age Group
(years)
Table 2. Percentage of Deaths During 16 Years According
Deaths
Participants
No.
Women
30-39
40-49
50-59
60-69
70-79
801
1692
2773
1634
370
36
167
561
Total 30-79
7270
1656
30-39
40-49
50-59
60-69
70-79
733
1427
1991
1554
384
47
247
712
954
323
Total 30-79
6089
2283
638
254
tO Grayinp nf thp. Hair hv SPY and AwfI*
%
4.5
9.9
20.2
39.1
68.7
Men
6.4
17.3
35.8
61.4
84.1
Age Group
(years)
Women
30-39
40-49
50-59
60-69
70-79
Men
30-39
40-49
50-59
60-69
70-79
No Gray
Hair
Few Gray
Hairs
Moderate
Gray Hairs
Completely Gray/
White Hairs
4.9
10
17
30
—
4.8
10
22
39
69
8.1
20
41
64
7.8
21
6.1
6.8
18
37
63
81
13
30
56
*(— means < 20 oersoris in category'1.
—
17
34
61
85
41
72
—
18
38
62
85
SIGNS OF AGING AND LONGEVITY
Wrinkles.—Table 5 shows the extent of facial wrinkles
(crow's feet) and the mortality by sex and age-groups.
For both sexes, the Cox regression analysis showed no
significant correlation between facial wrinkles and mortality.
Arcus senilis.—Table 6 shows the extent of arcus senilis
and the mortality by sex and age groups. For women, we
found a significant correlation between arcus senilis and
mortality (p < .01). In comparison with women without an
arcus senilis, the relative risk among those with a half arcus
was 1.13; 95% CI 1.00-1.46 and with a complete arcus
1.25; 95% CI 1.08-1.46.
In men, the same tendency was observed, although the
correlation was not statistically significant (p = .15).
DISCUSSION
In general, we found no correlation between the extent of
graying of the hair, baldness, facial wrinkles, and longevity
in a large random sample of men and women 30-79 years
of age living in Copenhagen. Thus, persons who look older
than their contemporaries seem to have the same life span.
The only exception to this conclusion regards a small fraction of men without any gray hair, who seem to have a significantly lower mortality.
In a previous analysis of the same three aging variables,
we found that, in addition to established coronary risk factors, graying of the hair, baldness, and facial wrinkles were
significantly correlated to the risk of myocardial infarction
in men, but not in women (1).
Lasker and Kaplan (9) have found an association of
gray hair and mortality in 480 Mexicans examined in 1948;
they stated, however, that "the observed relationship may
not generally be true." Our study could not confirm their
findings.
With regard to baldness, Herrera et al. (2), reporting data
from the Framingham Study, found no correlation between
the extent of baldness in men and all-cause mortality. This
is in line with our results. However, in the Framingham
Study the extent of baldness was assessed twice with 6
years apart. When the progression of baldness was analyzed
with respect to all-cause mortality, men with a rapid pro-
Table 5. Percentage of Deaths During 16 Years According
to Facial Wrinkles (Crow's foot)*
Table 3. Percentage of Deaths During 16 Years According
to Frontopariental Baldness by Sex and Age*
Age Group
(years)
No Bald
Triangle
Small Bald
Triangle
Large Bald
Triangle
4.5
9.4
20
38
67
3.2
16
20
45
67
—
—
—
—
—
7.2
18
38
63
84
5.3
16
33
61
84
Women
30-39
40-49
50-59
60-69
70-79
Men
Age
Group
(years)
Women
30-39
40-49
50-59
60-69
70-79
No
Wrinkles
(1)
Shallow
Wrinkles
(2)
Shallow
Wrinkles
(3)
Prominent
Wrinkles
(4)
Prominent
Wrinkles
(5)
3.8
4.5
8.3
21
37
—
4.9
9.9
20
36
63
4.5
10.6
19
40
67
28
39
77
7.3
14
35
4.9
16
34
56
—
65
8.8
18
36
61
86
7.7
23
—
—
Men
30-39
40-49
50-59
60-69
70-79
11
M349
5.8
17
.
43
58
87
(— means < 20 persons in category).
30-39
40-49
50-59
60-69
70-79
Women
30-39
40-49
50-59
60-69
70-79
Men
30-39
40-49
50-59
60-69
70-79
Thick
Hair
73
—
22
39
61
84
*(— means < 20 persons in category).
Table 4. Percentage of Deaths During 16 Years According
to Crown-top Baldness*
Age Group
(years)
6.2
25
—
—
—
Partly
Thin Hair
Bald
Spot
Bald Top
and Front
—
—
—
—
—
—
—
—
Table 6. Percentage of Deaths During 16 Years According
to the Extent of Arcus Senilis*
Age Group
(years)
No
Arcus
Half
Arcus
Complete
Arcus
4.2
9.5
19
37
65
—
7.8
26
40
71
—
35
26
47
72
—
—
17
Women
4.5
9
20
39
66
46
82
6.6
17
4.9
19
36
62
87
37
63
81
—
—
—
•
—
30-39
40-49 .
50-59
60-69
70-79
Men
*(— means < 20 persons in category).
2.8
18
38
—
12
61
88
60
83
30
30-39
40-49
50-59
60-69
70-79
6.2
17
34
59
86
*(— means < 20 persons in category).
23
41
59
82
37
68
84
M350
SCHNOHR ETAL.
gression were found to have a 2.4 higher risk of death than
those with slight or moderate progression.
We found a significant correlation between arcus senilis
and longevity, but only in women. A similar but not significant tendency was found among men. Several studies have
shown an association between arcus senilis and ischemic
heart disease. Chambless et al. (10) have found that in men
under 50 years of age, arcus senilis is a predictor of coronary heart disease, whereas no conclusions could be drawn
for women. Thus, the available data suggest that arcus
senilis is a predictor of coronary heart disease and all-cause
mortality, although we found the latter correlation only significant in women. We have no good explanation for the
discrepancy between the sexes and suggest further studies.
In recent years there has been an increasing interest in
the aging process, particularly in factors that may postpone
it. Although model diseases of premature aging such as
Werners syndrome (11) have been known for decades, only
recently the genes involved have been identified (12). We
had expected to find a correlation between the degree of
hair graying, baldness, facial wrinkles, and longevity, but
we were not able to demonstrate this in our analysis based
on cross-sectional assessment. If further studies based on
follow-up examinations demonstrate that persons with a
rapid progression of aging signs are prone to a shorter lifespan, then we have to seek for environmental and genetic
factors that eventually could be modified and thus prevent
premature death.
Heart Study, Bispebjerg Hospital-Entrance 33, DK-2400 Copenhagen
NV, Denmark. E-mail: [email protected]
REFERENCES
1. Schnohr P, Lange P, Nyboe J, Jensen G. Gray hair, baldness, and
wrinkles in relation to myocardial infarction: The Copenhagen City
Heart Study. Am Heart J. 1995;130:1003-1010.
2. Herrera CR, D'Agostino RB, Gerstman BB, Bosco LA, Belanger AJ.
Baldness and coronary heart disease rates in men from the Framingham study. Am J Epidemiol. 1995; 142:828-833.
3. Appleyard M, Hansen AT, Schnohr P, Jensen G, Nyboe J. The Copenhagen City Heart Study. A book of tables with data from the first
examination (1976-78) and a five-year follow-up (1981-83). Scand J
SocMed. 1989;170(suppl41):l-160.
4. Rose GA, Blackburn H. Cardiovascular survey methods. Geneva:
World Health Org Monogr. 1968;Ser 56.
5. Hamilton JB. Patterned loss of hair in man: types and incidence. NY
AcadSci. 1950;53:708-728.
6. Daniell HW. Smoker's wrinkles. Ann Intern Med. 1971 ;75:873-880.
7. Cox DR. Regression models and life tables. J R Stat Soc (B).
1972;35:187-220.
8. Gordon T. Hazards in the use of logistic function with special reference to data from prospective cardiovascular studies. J Chron Dis.
1974;27:97-102.
9. Lasker GW, Kaplan B. Graying of the hair and mortality. Soc Biol.
1970;21:290-295.
10. Chambless LE, Fuchs FD, Linn S, et al. The association of corneal
arcus with coronary heart disease and cardiovascular disease mortality in the Lipid Research Clinics mortality follow-up study. Am J
Public Health. 1990;80:1200-1204.
11. Epstein CJ, Martin GM, Schultz AL, Motulsky AG. Werner's syndrome. Medicine. 1966;45:177-221.
12. Goto M, Rubenstein M, Weber J, Woods K, Drayana D. Genetic linkage of Werner's syndrome to five markers of chromosome 8. (Letter
to the editor.) Nature. 1992,335:735-738.
ACKNOWLEDGMENTS
This study was supported by grants from The Danish Heart Foundation.
Address correspondence to Dr. Peter Schnohr, The Copenhagen City
Received December 11, 1996
Accepted December 9, 1997