Sie sind auf Seite 1von 8

Vol. 161, No.

12
Printed in U.S.A.
DOI: 10.1093/aje/kwi147

American Journal of Epidemiology


Copyright 2005 by the Johns Hopkins Bloomberg School of Public Health
All rights reserved

ORIGINAL CONTRIBUTIONS
Acne in Adolescence and Cause-specic Mortality: Lower Coronary Heart
Disease but Higher Prostate Cancer Mortality
The Glasgow Alumni Cohort Study

B. Galobardes1, G. Davey Smith1, M. Jeffreys2, S. Kinra1, and P. McCarron3


1

Received for publication August 16, 2004; accepted for publication January 27, 2005.

Androgen level or androgen activity is implicated in several health outcomes, but its independent role remains
controversial. This study investigated the association between history of acne in young adulthood, a marker of
hormone activity, and cause-specic mortality in the Glasgow Alumni Cohort Study. Male students who attended
Glasgow University between 1948 and 1968 and participated in voluntary health checks reported history of acne
(n 11,232). Vital status has been traced, and risk factors in adulthood are known for about 50% of the
participants. Those with a history of acne were more often nonsmokers while university students and tended
to be from a lower socioeconomic position. The two groups did not differ in other adolescent (height, body mass
index, blood pressure, and number of siblings) or in most adult risk factors. Students who reported a history of
acne had a lower risk of all-cause (hazard ratio 0.89, 95% condence interval (CI): 0.76, 1.04) and coronary
heart disease (hazard ratio 0.67, 95% CI: 0.48, 0.94) mortality but had some evidence of a higher risk of
prostate cancer mortality (hazard ratio 1.67, 95% CI: 0.79, 3.55). This study shows that androgen activity
during adolescence may protect against coronary heart disease but confer a higher risk of prostate cancer
mortality.
acne vulgaris; androgens; coronary disease; hormones; mortality; prostatic neoplasms

Abbreviations: CI, condence interval; HR, hazard ratio; ICD-9, International Classication of Diseases, Ninth Revision; ICD-10,
International Classication of Diseases, Tenth Revision.

Editors note: An invited commentary on this article


appears on page 1102.

heart disease risk, while acknowledging that firm conclusions


cannot be drawn from the currently available evidence (2).
Observational studies have examined the association between
markers of hormonal status and disease. Men who suffered
myocardial infarction shaved less frequently, an indication of
low androgen activity, than did a group of controls (3). In
a recent prospective study, Ebrahim et al. (4) reported higher
all-cause, coronary heart disease, and stroke mortality among
men who shaved less frequently. On the other hand, hair loss,

Testosterone has been implicated in prostate cancer development and, although once thought to explain the higher risk
of coronary heart disease in men, its role in health outcomes
is controversial (1). A recent review concluded that circulating androgens have a neutral or protective effect on coronary

Correspondence to Dr. Bruna Galobardes, Department of Social Medicine, University of Bristol, Canynge Hall, Whiteladies Road, Bristol BS8
2PR, United Kingdom (e-mail: bruna.galobardes@bristol.ac.uk).

1094

Am J Epidemiol 2005;161:10941101

Downloaded from http://aje.oxfordjournals.org/ by guest on December 23, 2014

Department of Social Medicine, University of Bristol, Bristol, United Kingdom.


Centre for Public Health Research, Massey University, Wellington, New Zealand.
3
Department of Epidemiology and Public Health, The Queens University of Belfast, Belfast, United Kingdom.
2

Acne in Adolescence and Cause-specic Mortality

MATERIALS AND METHODS

Detailed information on the Glasgow Alumni Cohort is


available elsewhere (10). Briefly, between 1948 and 1968,
students at Glasgow University were invited to attend
a health examination carried out by physicians. Information
on sociodemographic characteristics, health behaviors, and
medical history, including history of acne, was obtained by
means of questionnaire. Data collected during the physical
examination included measurements of height, weight, and
blood pressure. A total of 11,756 men, representing about 50
percent of the complete male student population, participated in the study. Since 1998, approximately 85 percent
(n 9,919) of the male cohort has been successfully traced
through the National Health Service Central Register, which
provides continuous updates on the date and cause of death
for members of the cohort. There were 1,315 deaths up to
February 2004. During 2001 and 2002, members of the cohort that were still alive were contacted through a postal
questionnaire (n 8,410) that sought to determine additional childhood information and adult sociodemographic
characteristics and risk factors. About 50 percent of
the male cohort (n 4,191) responded to the adult postal
Am J Epidemiol 2005;161:10941101

follow-up. The authors obtained ethical approval for the


study and informed consent from the participants.
Only men are included in this report because of the small
number of women attending the university from 1948 to
1968 and the low number of female deaths. Participants with
missing information on acne history were excluded (n
39), as were students aged more than 30 years at the time
of examination (n 479) and those with unknown date of
death (n 6), resulting in a total of 11,232 men included in
these analyses.
Variable description

History of acne was ascertained during the students visit


to the health services of the university. It was noted in the
medical history section of the questionnaire, but details
were not available on how the history of acne was measured
or on any treatment received. Given the prevalence of history of acne in this cohort, we hypothesize that mainly severe acne was recorded (11). Childhood socioeconomic
position was assigned by coding the fathers occupation into
social class, a five-point scale from I (most affluent) to V
(least affluent), using the Registrar Generals classification
(12, 13). Age (years), number of siblings, height (meters),
body mass index (kg/m2), systolic and diastolic blood pressure (mmHg), current smoking habit, occasional and regular
alcohol consumption, and history of acne were recorded
while the student was at the university.
Among those who participated in the postal follow-up in
20012002, the following adult characteristics were obtained: adult socioeconomic position based on the main
occupation held (from I to V), height (meters), body mass
index (kg/m2), marital status, smoking (never, former, or
current), physical activity (yes/no to a question about performing sufficient exercise to work up a sweat) (14), alcohol
consumption in the last 12 months (12 days/week vs. 3
days per week), and physicians diagnosis of hypertension
(present/absent).
Cause-specic mortality

International Classification of Diseases, Ninth Revision


(ICD-9) and Tenth Revision (ICD-10), codes were used to
group cause-specific mortality: coronary heart disease (ICD-9
codes 410414 and 429.2; ICD-10 codes I20I25 and I51.6);
stroke (ICD-9 codes 431438; ICD-10 codes I61I69 and
G45); lung cancer (ICD-9 code 162; ICD-10 code C34);
prostate cancer (ICD-9 code 185; ICD-10 code C61); colon
cancer (ICD-9 code 153; ICD-10 code C18); and external
causes of death including accidents, suicide, and violence
(ICD-9 codes 800999 and E800E999; ICD-10 codes
S00T98 and V01Y89).
Statistical analysis

Cox proportional hazards models were used to estimate the


risk of overall and cause-specific mortality associated with
history of acne in young adulthood, adjusting for examination
date, socioeconomic position, and several disease risk factors
measured at the university, which are potential confounders
of the association between acne and cause-specific mortality.

Downloaded from http://aje.oxfordjournals.org/ by guest on December 23, 2014

a marker of greater androgen activity, has been associated


with a higher risk of coronary heart disease.
Acne vulgaris first develops at the onset of puberty as a
result of hormonal changes, although additional mechanisms
are also involved (6). The initial lesion, the comedo, fills
with lipids because of increased activity of the sebaceous
gland and becomes visible under androgen stimulation. Inflammatory acne results from the action of Propionibacterium acnes, a pathogen present in normal skin flora which
metabolizes sebaceous triglycerides, consuming glycerol
and releasing free fatty acids, neutrophil, and complement
attractants (7). Results from a twin study (8) and a study
demonstrating increased risk of acne in first-degree relatives
of adults with persistent acne (9) suggest that acne is partially inherited. No major environmental exposure has been
consistently reported. Acne can be a marker of increased
androgen activity during puberty and, if acne risk is related
to health outcomes, this would support a potential role of
hormones in disease etiology.
Androgens are the main hormones involved in the development of acne, although the specific mechanisms are not
known. In addition, acne is rare in adulthood despite androgen
levels remaining high. Other hormones, such as growth hormone and insulin-like growth factors, increase in puberty, raising the possibility that they contribute to acne development,
and they are also considered to be linked to adult disease.
To our knowledge, the association between history of acne
in young adulthood and mortality has not been studied. We
investigated the hypothesis that a marker of greater androgen
activity during adolescence may affect subsequent risk of
chronic disease in a cohort of former students who underwent
clinical examinations at Glasgow University between 1948
and 1968 and for whom a history of acne and information on
cause-specific mortality and selected disease risk factors are
available.

1095

1096 Galobardes et al.

TABLE 1. Age-adjusted means and prevalence of childhood and young adulthood


characteristics,* by history of acne, Glasgow Alumni Cohort Study, 19481968 (baseline)
History of acne
(mean (SDy))

No history of acne
(mean (SD))

p value

Age (years) (n 11,232)

20.0 (2.40)

20.7 (2.81)

Height (cm) (n 11,196)

174.96 (6.37)

174.78 (6.35)

0.255

Body mass index (kg/m )


(n 11,189)

21.57 (2.23)

21.60 (2.23)

0.549

Systolic blood pressure


(mmHg) (n 11,182)

131.06 (13.01)

130.58 (12.97)

0.135

<0.001

Diastolic blood pressure


(mmHg) (n 11,159)

77.02 (8.52)

77.36 (8.49)

0.114

No. of siblings (n 11,229)

1.71 (1.56)

1.72 (1.56)

0.723

No.

No.

1,948

53.25

8,882

56.86

0.004

No smoking (n 10,703)

1,943

68.20

8,760

65.59

0.022

Occasional or regular alcohol


consumption (n 9,584)

1,798

55.30

7,786

54.31

0.446

* Age-adjusted values for all variables except age.


y SD, standard deviation.

It was not possible to adjust for adult risk factors, because to


date there have been few deaths in those who participated in
the 20012002 follow-up and for whom adult characteristics
are known, but we present a descriptive analysis of this subgroup of former students.

RESULTS

A history of acne was reported by 2,024 students (18.0


percent). Age, height, body mass index, systolic and dia-

stolic blood pressure, and number of siblings at the university were similar among those with and those without a
history of acne (table 1). Students with a history of acne
were more often nonsmokers at the university (68.20 percent vs. 65.59 percent; p 0.022) and from a lower socioeconomic background. In adulthood, those with a history of
acne were more likely to have never smoked (47.74 percent
vs. 43.34 percent; p 0.027) (table 2). There were no differences in adult social class, marital status, height, body mass
index, physical activity, alcohol consumption, fat consumption, or hypertension.

TABLE 2. Age-adjusted means and prevalence of descriptive adulthood characteristics


in the subgroup of participants that responded to a postal follow-up, by history of acne,
Glasgow Alumni Cohort Study, 20012002

Height (cm) (n 4,031)

History of acne
(mean (SD*))

No history of acne
(mean (SD))

p value

176.41 (6.52)

176.47 (6.50)

0.809

25.52 (3.13)

0.883

Body mass index (kg/m )


(n 3,961)

25.50 (3.14)
No.

No.

Married or living together


(n 4,044)

770

90.56

3,274

88.32

0.061

Adult social class I


(n 2,971)

578

51.28

2,393

52.48

0.606

Never smoked (n 4,058)

774

47.74

3,284

43.34

0.027

No physical activity
(n 4,044)

774

33.82

3,270

34.33

0.786

Hypertension (n 4,010)

765

30.54

3,245

31.00

0.801

Alcohol consumption on
12 days/week or less
(n 4,018)

763

38.92

3,255

36.96

0.319

* SD, standard deviation.

Am J Epidemiol 2005;161:10941101

Downloaded from http://aje.oxfordjournals.org/ by guest on December 23, 2014

Fathers social class III


(n 10,830)

Acne in Adolescence and Cause-specic Mortality

1097

TABLE 3. Hazard ratio and 95% condence interval of all-cause and cause-specic mortality among men with a history of acne
compared with those who did not have acne, Glasgow Alumni Cohort Study, 19482004
No. of
deaths

Hazard
ratio

95% condence
interval

1,315

0.89

0.76, 1.03

Cardiovascular disease

527

0.74

Coronary heart disease

369

0.67

Stroke

88

Lung cancer

No. of
deaths

Hazard
ratio

95% condence
interval

1,286

0.87

0.74, 1.01

0.57, 0.95

518

0.72

0.49, 0.93

362

0.65

1.27

0.75, 2.15

88

87

1.25

0.73, 2.13

Prostate cancer

43

1.58

Colon cancer

47

1.03

External causes

79

1.05

All causes

Model 3z

Model 2y

Model 1*
Causes of death

No. of
deaths

Hazard
ratio

95% condence
interval

1,213

0.89

0.76, 1.04

0.55, 0.94

482

0.74

0.56, 0.96

0.47, 0.90

336

0.67

0.48, 0.94

1.26

0.74, 2.14

81

1.24

0.71, 2.17

83

1.22

0.71, 2.11

80

1.27

0.72, 2.24

0.78, 3.20

41

1.48

0.71, 3.12

37

1.67

0.79, 3.55

0.48, 2.21

47

1.02

0.48, 2.19

43

1.01

0.45, 2.27

0.60, 1.85

76

1.02

0.57, 1.82

72

1.08

0.60, 1.94

Men with a history of acne measured at the university had


lower overall (hazard ratio (HR) 0.89, 95 percent confidence interval (CI): 0.76, 1.04), cardiovascular (HR 0.74,
95 percent CI: 0.56, 0.96), and coronary heart disease
(HR 0.67, 95 percent CI: 0.48, 0.94) mortality (table 3).
There was some evidence that these men had a higher risk of
death from prostate cancer (HR 1.67, 95 percent CI: 0.79,
3.55). History of acne was not associated with mortality due
to stroke, lung cancer, colon cancer, or external causes.
Adjustment for childhood socioeconomic circumstances
and for behavioral and biologic risk factors present while
the student attended the university did not affect these estimates. Excluding people with missing confounders did not
change the estimates in models 1 and 2.
To further control for potential residual confounding due
to a higher proportion of nonsmokers among those with
a history of acne, we investigated the mortality pattern
among the students who reported being nonsmokers at the
university. For those reporting a history of acne compared
with those who did not, the fully adjusted hazard ratio was
0.82 (95 percent CI: 0.66, 1.02) for all causes of death, 0.68
(95 percent CI: 0.46, 0.99) for cardiovascular disease, 0.65
(95 percent CI: 0.41, 1.02) for coronary heart disease, and
2.17 (95 percent CI: 0.94, 5.01) for prostate cancer mortality.

DISCUSSION

A history of acne in young adulthood was associated with


lower overall, cardiovascular disease, and coronary heart
disease mortality but with higher prostate cancer mortality.
This mortality pattern was also observed when analyses
were restricted to nonsmokers. Adjustment for several socioeconomic and disease risk factors measured at the university did not affect the associations. In addition, the
observed pattern of cause-specific mortality suggests that
confounding is unlikely to explain these results.
Am J Epidemiol 2005;161:10941101

Androgen activity

Several hormones are implicated in the development of


acne and, although androgens play a key role in this process,
the specific mechanisms involved are not well understood.
Administering androgens produces acne, and castrated men
or men with genetic mutations that affect androgen metabolism or receptor function do not develop acne (15). Women
with acne benefit from antiandrogen therapy (16) and from
oral contraceptive therapy, which is believed to reduce androgen metabolites (17). However, it is not clear why acne
spontaneously improves and disappears in most people.
While acne recedes in late adolescence, androgen levels
remain high until late adulthood, indicating that androgens alone cannot be responsible for the presence of acne
(18, 19). Given the prevalence of history of acne among
the Glasgow Alumni Cohort, we hypothesized that mainly
severe cases of acne were recorded. However, androgen
levels do not always correlate with acne severity (20) and,
in studies reporting higher hormonal levels in acne patients,
these are nevertheless within the normal physiologic range
(21). In addition, many men do not develop noticeable levels
of acne despite their high levels of androgens compared with
women. Thus, whether it is only severe acne that relates to
adult mortality risk cannot be assessed in our study.
Other markers of androgen activity have been previously
related to health outcomes. Infrequent shaving, indicative of
low androgen levels or of low androgen activity, is related to
higher coronary heart disease and stroke risk (4). The sensitivity of sexual hair follicles to androgens depends on their
distribution. Tenfold higher androgen concentrations are
needed for the beard to appear compared with pubic hair,
although there is great individual variability in the actual
level of androgens (19). Greater sexual activity, another
marker of higher androgen concentrations, was found to
protect from coronary heart disease and, less clearly, from
stroke in the Caerphilly Study (22). These sex-related
characteristicsdevelopment of facial hair and frequency

Downloaded from http://aje.oxfordjournals.org/ by guest on December 23, 2014

* Model 1: adjusted for date of examination at Glasgow University Health Service.


y Model 2: adjusted for date of examination, fathers socioeconomic position, and number of siblings.
z Model 3: adjusted for date of examination, fathers socioeconomic position, number of siblings, height, body mass index, cigarette consumption, and systolic blood pressure.

1098 Galobardes et al.

in puberty, such as an early growth spurt and an early age at


first ejaculation, showed a higher risk (36). The role of
androgens on the prostate gland has not been elucidated,
and whether they induce prostate cancer or facilitate the
growth of existing lesions is not clear (26). During puberty,
prostate-specific antigen levels increase, and prostate epithelial differentiation occurs (37), indicating that changes in
adolescence may modify later risk of prostate cancer. A
review of prospective studies showed no compelling evidence that testosterone levels are causally related to prostate
cancer (1). A meta-analysis reported a higher risk of prostate
cancer with higher levels of sexual activity, a marker of
higher testosterone levels (38), whereas Leitzmann et al.
(37) found that a high ejaculation frequency was associated
with lower, not higher, risk of total and localized prostate
cancer. Vertex baldness has also been associated with higher
prostate cancer risk (39).
Other hormonal changes in acne

Growth hormone and insulin-like growth factors undergo


dramatic changes during puberty. Growth hormone rises
during pubertal development because of increases in gonadal sex steroids in boys and girls, but its secretion decreases after puberty (40). Similarly, insulin-like growth
factor levels rise during puberty and decrease 12 years
thereafter. This pattern corresponds closely with that of acne
development and remission, but the role of these hormones
in acne production remains largely unknown (19). Growth
hormone seems to facilitate the effect of androgens on pubic
hair growth; that is, the testosterone level required to induce
pubic hair in growth hormone-deficient boys is much higher
than that needed by boys with normal growth hormone levels (19). Some of the androgen effect on hair growth may be
mediated through insulin-like growth factors, since these
factors have been found to induce important alterations in
the hair growth cycle (19). Given that achieved adult height
is associated with an increased risk of prostate cancer (41)
but a lower risk of cardiovascular disease (4245) and type
II diabetes (46), growth hormone levels during puberty
could account for both the protective effect of a history of
acne on coronary heart disease and the higher risk of prostate cancer. However, in our study, height measured at the
university or self-reported in adulthood was similar in students who reported a history of acne and those who did not,
providing no support for this hypothesis. Moreover, growth
during early life is mainly due to increases in leg length,
while postpubertal growth occurs through increases in both
leg and trunk length (47). The association of greater height
with increased cancer risk and lower cardiovascular disease
and type II diabetes risk seems to be due to increased leg
length, suggesting that the height-related exposures are
more likely to occur before puberty (41, 46, 48, 49).
Confounding

Although we adjusted for available risk factors, confounding may still remain an issue in observational studies
(50). Adjustment for more detailed categorizations of
Am J Epidemiol 2005;161:10941101

Downloaded from http://aje.oxfordjournals.org/ by guest on December 23, 2014

of intercourseand acne have a common androgen-related


basis initiated at puberty.
Baldness, in particular of the male pattern (i.e., vertex but
not frontal baldness) (23), baldness occurring at a younger
age (24), and progression of baldness (25) are markers of
higher androgen activity but, unlike other markers of androgen activity, are associated with higher risk of coronary heart
disease (5, 2325). The paradoxical effects of higher androgen activity in the form of hair loss (baldness) and of higher
facial hair (beard growth) point to local or tissue-specific
differences in androgen effects, which could also be relevant
for the association of androgens with health outcomes.
The role of androgens with regard to health is not established, but debate about the potential benefits and risks of
androgen replacement therapy in aging men is increasingly
found in the medical literature. Clinical trials, particularly of
long exposure to androgens, are rare (26). The Committee
on Assessing the Need for Clinical Trials of Testosterone
Replacement Therapy concluded that there is no sufficient
evidence that testosterone treatment is beneficial in elderly
men with symptoms that might be caused by hypogonadism
(27). In a recent review, the authors reported conflicting
results within studies and found that most epidemiologic
studies lacked control for potential confounders (2). Among
the studies that attempted to control for confounders, testosterone and dihydrotestosterone appeared to have a protective
effect on cardiovascular disease (2). Another review found
a potential coronary heart disease benefit from high levels of
androgens (28). A case-control study reported lower levels
of androgens among cases of coronary artery disease compared with controls, although there was no dose relation
between androgen levels and severity of coronary disease
(29). The Rotterdam Study reported an inverse association
between free testosterone levels and aortic atherosclerosis
among men who were nonsmokers (30). Low levels of
testosterone seem to correlate with higher levels of atheroma, mediated by coronary artery vasodilatation and vascular inflammation (31), and with a lower ejection fraction of
the left ventricle in males (32). However, in a twin study, no
correlation was found between endogenous sex hormones
and future coronary heart disease risk (33). The cardioprotection conferred by testosterone could be through effects on
coronary heart disease risk factors (26, 30), but there is
conflicting evidence on its role in the etiology of insulin
resistance and hypertension (2). On the other hand, a genetic
polymorphism, shorter polyglutamine stretches in exon 1 of
the androgen receptor gene (CAG repeats) determining
higher androgen activity, was associated with more severe
coronary artery stenosis (34). In some studies, androgen
exposure has been associated with higher, not lower, risk.
High levels of exogenous androgens, such as found in some
transsexual females and both male and female athletes, are
associated with impaired endothelial vasodilatation and premature myocardial infarction and stroke (2, 35), although
most of this evidence is from case series.
There was some evidence in our study that students with
acne history have a higher risk of prostate cancer mortality.
A case-control study in Australia found a protective role of
severe acne (presence of facial scarring in adulthood) on
prostate cancer, whereas other markers of androgen action

Acne in Adolescence and Cause-specic Mortality

Am J Epidemiol 2005;161:10941101

members of the Glasgow Alumni Cohort were more affluent


than those of the general population: in 19511961, 53.8
percent of the Glasgow Alumni Cohort were from social
classes I and II compared with 6.6 percent of similarly aged
individuals in the general Scottish population (10).
However, it seems unlikely that the influence of androgen
activity on coronary heart disease risk should differ among
men of low and high social class. In addition, associations
reported from this cohort between early life exposures and
later disease will be less likely to have resulted from
confounding by adult socioeconomic circumstances, given
the relative homogeneity of adulthood social circumstances
in this group.
Studies examining the role of exposure to endogenous
hormone levels or hormone activity in adult life in coronary
heart disease or prostate cancer etiology have not been
conclusive. We have shown that an early life measure related to puberty development was associated with these outcomes. Other studies assessing early life exposure to
hormone activity may provide a better understanding of
the role of hormones. In addition, studies based on principles of Mendelian randomization, using genetic polymorphisms related to hormone levels or hormone activity, are
more likely to provide unconfounded estimates of disease
risks associated with endogenous hormonal exposures (61).
Such studies can assist in establishing the role of early life
influences in coronary heart disease and prostate cancer and
in providing insight into the mechanisms that may explain
these associations, which, in turn, can lead to the development of preventative interventions.

Conclusion

Higher androgen levels or higher androgen activity could


plausibly explain the observed lower coronary heart disease
and higher prostate cancer mortality among men with a
history of acne. Confounding was unlikely to explain these
results. The effects of androgenic hormone activity in early
life warrant more research with alternative study designs.

ACKNOWLEDGMENTS

The authors would like to acknowledge the financial


support of the Stroke Association; Chest, Heart and Stroke
Scotland; the National Health Service Research and Development Cardiovascular Disease Programme; and the
World Cancer Research Fund. G. D. S. holds a Robert Wood
Johnson Foundation Investigators Award in Health Policy
Research. Funds from this award partly supported B. G.
In addition, P. M. is supported by a career scientist
award funded by the Research and Development Office for
Health and Personal Social Services in Northern Ireland.
The Centre for Public Health Research (Massey University,
Wellington, New Zealand) is supported by a Programme
Grant from the Health Research Council of New Zealand.
The authors work was independent of the funding
sources.

Downloaded from http://aje.oxfordjournals.org/ by guest on December 23, 2014

potential confounders did not change the magnitude of the


associations. A higher proportion of students with a history
of acne did not smoke while at the university compared with
those that reported no history of acne. In adulthood, men
with a history of acne reported more often to have never
smoked, but we could not account for potential adulthood
confounders in the mortality analysis.
The pattern of cause-specific mortality risk observed in
our study allows us to evaluate whether confounding is
a plausible explanation for these results, as follows. Lower
socioeconomic circumstances during childhood are associated with higher coronary heart disease and stroke risk (51),
results also found in this cohort (52). Therefore, poorer
childhood circumstances among those with a history of acne
cannot explain a lower coronary heart disease risk. If a lower
prevalence of smoking among students with a history of
acne was responsible for the observed lower coronary heart
disease mortality, then, given the stronger association of
smoking with lung cancer, we should have also observed
lower lung cancer mortality, which was not the case. In
addition, the association between history of acne and coronary heart disease mortality was similar among students
who were nonsmokers at the university to that of the whole
cohort.
Another potential confounder to consider is diet. Those
reporting a higher prevalence of history of acne were of
a lower childhood socioeconomic position (although still
higher than that of the population that did not attend the
university) and could therefore have been exposed to a less
healthy diet. This, however, cannot explain the lower coronary heart disease risk in men who had acne in earlier life.
On the other hand, fat restriction might have been prescribed
to those students who had acne (mainly carbohydrate and
also fat restriction were prescribed as treatment for acne at
that time) (53, 54). This could potentially explain the lower
coronary heart disease risk later in life. However, the association of total fat intake or specific dietary fats with prostate cancer risk is not established (55) and higher, not lower,
fat intake would be associated with higher risk of prostate
cancer mortality.
Progressively after the 1950s, tetracyclines were used for
the treatment of acne (56, 57). Infectious contributions, including from Chlamydia pneumoniae, to coronary heart disease risk have been proposed, although these remain
controversial (58, 59). However, if C. pneumoniae infection
is causally related to coronary heart disease, men who had
their acne treated with antibiotics could consequently be at
lower coronary heart disease risk. On the other hand, tetracycline or its derivatives have been shown to inhibit cell
growth in prostate cancer (60), and therefore we would
not expect a higher risk of prostate cancer mortality in the
acne group.
Finally, mortality due to external causes (accidents, suicide, and violence) was not related to history of acne,
whereas if acne relates to coronary heart disease mortality
because of behavioral confounders, a relatively low mortality from external causesunlikely to be related to androgen
levels or activitywould also be expected.
Overall, this evidence suggests that confounding does not
explain the pattern of mortality observed in our study. The

1099

1100 Galobardes et al.

REFERENCES
23.
24.
25.
26.
27.
28.
29.
30.

31.

32.

33.
34.

35.
36.
37.
38.
39.
40.

41.
42.

Am J Epidemiol 2005;161:10941101

Downloaded from http://aje.oxfordjournals.org/ by guest on December 23, 2014

1. Rhoden EL, Morgentaler A. Medical progressrisks of


testosterone-replacement therapy and recommendations for
monitoring. N Engl J Med 2004;350:48292.
2. Muller M, van der Schouw YT, Thijssen JHH, et al. Endogenous sex hormones and cardiovascular disease in men. J Clin
Endocrinol Metab 2003;88:507686.
3. Phillips GB. Evidence for hyperoestrogenaemia as a risk factor
for myocardial infarction in men. Lancet 1976;2:1418.
4. Ebrahim S, Davey Smith G, May M, et al. Shaving, coronary
heart disease, and stroke: the Caerphilly Study. Am J Epidemiol 2003;157:2348.
5. Herrera CR, Lynch C. Is baldness a risk factor for coronary
artery disease? A review of the literature. J Clin Epidemiol
1990;43:125560.
6. Leyden JJ. New understandings of the pathogenesis of acne.
J Am Acad Dermatol 1995;32:S1525.
7. Webster GF, Poyner T, Cunliffe B. Acne vulgaris. Commentary: a UK primary care perspective on treating acne. BMJ
2002;325:4759.
8. Bataille V, Snieder H, MacGregor AJ, et al. The influence
of genetics and environmental factors in the pathogenesis
of acne: a twin study of acne in women. J Invest Dermatol
2002;119:131722.
9. Goulden V, McGeown CH, Cunliffe WJ. The familial risk of
adult acne: a comparison between first-degree relatives of
affected and unaffected individuals. Br J Dermatol 1999;141:
297300.
10. McCarron P, Davey Smith G, Okasha M, et al. Life course
exposure and later disease: a follow-up study based on medical
examinations carried out in Glasgow University (194868).
Public Health 1999;113:26571.
11. Galobardes B, Davey Smith G, Jeffreys M, et al. Has acne
increased? Prevalence of acne history among university students between 1948 and 1968. The Glasgow Alumni Cohort
Study. Br J Dermatol 2005;152:8245.
12. The Registrar-Generals decennial supplement, England and
Wales 1931. Part IIa. Occupational mortality. London, United
Kingdom: Her Majestys Stationery Office, 1931.
13. The Registrar-Generals decennial supplement, England and
Wales 1951. Part IIa. Occupational mortality. London, United
Kingdom: Her Majestys Stationery Office, 1951.
14. Godin G, Shephard RJ. A simple method to assess exercise
behavior in the community. Can J Appl Sport Sci 1985;10:
1416.
15. Shaw JC. Acne: effect of hormones on pathogenesis and
management. Am J Clin Dermatol 2002;3:5718.
16. Shaw JC. Antiandrogen and hormonal treatment of acne.
Dermatol Clin 1996;14:80311.
17. Carmina E, Godwin AJ, Stanczyk FZ, et al. The association
of serum androsterone glucuronide with inflammatory lesions
in women with adult acne. J Endocrinol Invest 2002;25:
7658.
18. Lucky AW. Hormonal correlates of acne and hirsutism. Am
J Med 1995;98:89S94S.
19. Deplewski D, Rosenfield RL. Role of hormones in pilosebaceous unit development. Endocr Rev 2000;21:36392.
20. Levell MJ, Cawood ML, Burke B, et al. Acne is not associated
with abnormal plasma androgens. Br J Dermatol 1989;120:
64954.
21. Thiboutot D, Gilliland K, Light J, et al. Androgen metabolism in sebaceous glands from subjects with and without
acne. Arch Dermatol 1999;135:10415.
22. Ebrahim S, May M, Ben Shlomo Y, et al. Sexual intercourse
and risk of ischaemic stroke and coronary heart disease: the

Caerphilly Study. J Epidemiol Community Health 2002;56:


99102.
Lesko SM, Rosenberg L, Shapiro S. A case-control study of
baldness in relation to myocardial infarction in men. JAMA
1993;269:9981003.
Ford ES, Freedman DS, Byers T. Baldness and ischemic heart
disease in a national sample of men. Am J Epidemiol 1996;
143:6517.
Herrera CR, DAgostino RB, Gerstman BB, et al. Baldness
and coronary heart disease rates in men from the Framingham
Study. Am J Epidemiol 1995;142:82833.
Alexandersen P, Christiansen C. The aging male: testosterone
deficiency and testosterone replacement. An up-date. Atherosclerosis 2004;173:15769.
Snyder PJ. Hypogonadism in elderly menwhat to do until
the evidence comes. N Engl J Med 2004;350:4402.
Alexandersen P, Haarbo J, Christiansen C. The relationship of
natural androgens to coronary heart disease in males: a review.
Atherosclerosis 1996;125:113.
English KM, Mandour O, Steeds RP, et al. Men with coronary
artery disease have lower levels of androgens than men with
normal coronary angiograms. Eur Heart J 2000;21:8904.
Hak AE, Witteman JCM, de Jong FH, et al. Low levels of
endogenous androgens increase the risk of atherosclerosis in
elderly men: the Rotterdam Study. J Clin Endocrinol Metab
2002;87:36329.
Malkin CJ, Pugh PJ, Jones RD, et al. Testosterone as a protective factor against atherosclerosisimmunomodulation
and influence upon plaque development and stability.
J Endocrinol 2003;178:37380.
Dobrzycki S, Serwatka W, Nadlewski S, et al. An assessment
of correlations between endogenous sex hormone levels and
the extensiveness of coronary heart disease and the ejection
fraction of the left ventricle in males. J Med Invest 2003;
50:1629.
Mikulec KH, Holloway L, Krasnow RE, et al. Relationship of
endogenous sex hormones to coronary heart disease: a twin
study. J Clin Endocrinol Metab 2004;89:12405.
Alevizaki M, Cimponeriu AT, Garofallaki M, et al. The androgen receptor gene CAG polymorphism is associated with
the severity of coronary artery disease in men. Clin Endocrinol
2003;59:74955.
Bagatell CJ, Bremner WJ. Androgens in menuses and
abuses. N Engl J Med 1996;334:70715.
Giles GG, Severi G, English DR, et al. Early growth, adult
body size and prostate cancer risk. Int J Cancer 2003;103:
2415.
Leitzmann MF, Platz EA, Stampfer MJ, et al. Ejaculation
frequency and subsequent risk of prostate cancer. JAMA
2004;291:157886.
Dennis LK, Dawson DV. Meta-analysis of measures of sexual
activity and prostate cancer. Epidemiology 2002;13:729.
Giles GG, Severi G, Sinclair R, et al. Androgenetic alopecia
and prostate cancer: findings from an Australian case-control
study. Cancer Epidemiol Biomarkers Prev 2002;11:54953.
Grumbach MM, Styne DM. Puberty: ontonegy, neuroendocrinology, physiology, and disorders. In: Larsen R,
Kronenberg HM, Melmed S, eds. Williams textobook of
endocrionology. 10th ed. Philadelphia, PA: Saunders,
Elsevier Science, 2004:1115286.
Gunnell D, Okasha M, Davey Smith G, et al. Height, leg
length, and cancer risk: a systematic review. Epidemiol Rev
2001;23:31342.
McCarron P, Okasha M, McEwen J, et al. Height in
young adulthood and risk of death from cardiorespiratory

Acne in Adolescence and Cause-specic Mortality

43.
44.

45.
46.

47.

49.

50.

51.

Am J Epidemiol 2005;161:10941101

52.

53.
54.
55.
56.
57.
58.

59.

60.
61.

adulthood: systematic review and interpretation. Epidemiol


Rev 2004;26:721.
Davey Smith G, McCarron P, Okasha M, et al. Social circumstances in childhood and cardiovascular disease mortality:
prospective observational study of Glasgow University students. J Epidemiol Community Health 2001;55:3401.
Borrie P. Roxburghs common skin diseases. 11th ed. London,
United Kingdom: HK Lewis and Co, Ltd, 1959.
Lewis GM. Practical dermatology. 2nd ed. Philadelphia, PA:
WB Saunders Company, 1959.
Dennis LK, Snetselaar LG, Smith BJ, et al. Problems with the
assessment of dietary fat in prostate cancer studies. Am J
Epidemiol 2004;160:43644.
Percival GH. An introduction to dermatology. 13th ed. Edinburgh, Scotland: E and S Livingstone, Ltd, 1967.
Rook A, Wilkinson DS, Ebling FJG. Textobook of dermatology. Vol II. 2nd ed. Oxford, United Kingdom: Blackwell Scientific Publications, 1972.
Axisa B, Loftus IM, Naylor AR, et al. Prospective, randomized, double-blind trial investigating the effect of doxycycline
on matrix metalloproteinase expression within atherosclerotic
carotid plaques. Stroke 2002;33:285864.
Baxter BT, Pearce WH, Waltke EA, et al. Prolonged administration of doxycycline in patients with small asymptomatic
abdominal aortic aneurysms: report of a prospective (phase II)
multicenter study. J Vasc Surg 2002;36:112.
Fife RS, Sledge GW Jr, Roth BJ, et al. Effects of doxycycline on human prostate cancer cells in vitro. Cancer Lett
1998;127:3741.
Davey Smith G, Ebrahim S. Mendelian randomization: can
genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 2003;32:
122.

Downloaded from http://aje.oxfordjournals.org/ by guest on December 23, 2014

48.

disease: a prospective study of male former students of Glasgow


University, Scotland. Am J Epidemiol 2002;155:6837.
Williams SR, Jones E, Bell W, et al. Body habitus and coronary heart disease in men. A review with reference to methods
of body habitus assessment. Eur Heart J 1997;18:37693.
Yarnell JW, Limb ES, Layzell JM, et al. Height: a risk marker
for ischaemic heart disease: prospective results from the
Caerphilly and Speedwell Heart Disease Studies. Eur Heart J
1992;13:16025.
Wannamethee SG, Shaper AG, Whincup PH, et al. Adult
height, stroke, and coronary heart disease. Am J Epidemiol
1998;148:106976.
Lawlor DA, Ebrahim S, Davey Smith G. The association between components of adult height and type II diabetes and
insulin resistance: British Womens Heart and Health Study.
Diabetologia 2002;45:1097106.
Gunnell D. Commentary: can adult anthropometry be used as
a biomarker for prenatal and childhood exposures? Int J
Epidemiol 2002;31:3904.
Gunnell DJ, Davey Smith G, Frankel S, et al. Childhood leg
length and adult mortality: follow up of the Carnegie (Boyd
Orr) Survey of Diet and Health in Pre-war Britain. J Epidemiol
Community Health 1998;52:14252.
Davey Smith G, Greenwood R, Gunnell D, et al. Leg length,
insulin resistance, and coronary heart disease risk: the
Caerphilly Study. J Epidemiol Community Health 2001;55:
86772.
Lawlor DA, Davey Smith G, Kundu D, et al. Those confounded vitamins: what can we learn from the differences
between observational versus randomised trial evidence?
Lancet 2004;363:17247.
Galobardes B, Lynch JW, Davey Smith G. Childhood socioeconomic circumstances and cause-specific mortality in

1101

Das könnte Ihnen auch gefallen