GENDER MEDICINE/VOL. 3, No.

2, 2 0 0 6

CSD Grand Rounds

Why Women Live Longer Than Men: Sex Differences in
Longevity
S t e v e n N. A u s t a d , P h D

Dr. Steve Austad received bachelor's degrees in literature from the University of California,
Los Angeles and in biology from California State University, followed by his PhD and post-
doctoral training in biological sciences at Purdue University. He joined the faculty at Harvard
University in the Department of Evolutionary Biology and rose through the ranks to Associate
Professor. Dr. Austad moved to the University of Idaho's Department of Biological Sciences in
1993 and became Full Professor in 1997. Currently, he is a professor in the Department of
Cellular and Structural Biology in the Barshop Institute for Longevity & Aging Studies at the
University of Texas Health Science Center in San Antonio. Dr. Austad has received the Robert
W. Kleemeier Award for Outstanding Research from the Gerontological Society of America
and the Nathan A. Shock Award from the National Institute of Aging. He has had a long his-
tory of grants from the National Institutes of Health, the National Science Foundation, and
private foundations, and has published extensively on aging and longevity. His book, W h y
We Age: W h a t Science Is Discovering About the Body's Journey Through Life, has been
translated into 7 languages and was named one of the 10 best books of 1998.

ABSTRACT
Historically, w o m e n have lived longer than m e n in almost every country in the world. A similar pat-
tern of sex differences in longevity is also found in m a n y other species; however, it is not clear if there
are more species in which females live longer or vice versa. For virtually all the primary causes of death
and at virtually all ages, mortality rates are higher for men. W o m e n do n o t live longer than m e n
because they age more slowly, but because they are more robust at every age. Paradoxically, although
w o m e n have lower mortality rates they have higher overall rates of physical illness than do men.
Several hypotheses have been proposed for sex differences in longevity, including more active female
i m m u n e functioning, the protective effect of estrogen, compensatory effects of the second X chromo-
some, reduction in the activity of growth h o r m o n e and the insulin-like growth factor 1 signaling cas-
cade, and the influence of oxidative stress on aging and disease. At present, none of these hypotheses
are strongly supported, although weak support is available for the oxidative stress hypothesis. With
the advent of more rapid genome sequencing, molecular tools will become available for more species,
thus further detailing the causes for the differences in longevity between the sexes.

INTRODUCTION
Some p h e n o m e n a , such as aging, are so ubiquitous we barely notice them. Few of us contemplate
w h y we age as opposed to w h y we don't. The health and longevity difference b e t w e e n the sexes is a

Based on a presentation to the Georgetown University Center for the Study of Sex Differences in Health, Aging and Disease
(CSD), October 20, 2005.
Accepted for publication March 31, 2006.
Gend Med. 2006;3:79-92. 1550-8579/06/$19.00

Copyright © 2006 ExcerptaMedica, Inc. 79

GENDERMEDICINE
similar p h e n o m e n o n ; although virtually every-
one can tell you t h a t w o m e n live longer t h a n
men, not m a n y people have asked why. There
is little research on this issue at a mechanistic
level. In this article, I will discuss h o w robust
these differences are between the sexes a n d h o w
we m i g h t elucidate t h e m using comparative
biology. I also will examine possible mechanis-
tic hypotheses and t h e n outline w h a t we are
starting to learn and w h y we d o n ' t k n o w more.
How Long Can Humans Live?
The oldest person that we k n o w of with a
valid birth record was Jeanne Calment, a French
w o m a n w h o lived to be 122 years old. To under-
stand exactly w h a t an accomplishment this
was, consider that this w o m a n was born w h e n
Ulysses S. Grant was president of the United
States, and that she died during Bill Clinton's
presidency. In the aging community, people
w h o have lived to be >110 years old are called
supercentenarians. Your chances of living to be
>110 years old are -1 in 3 million; thus, these
people are the elite of the elite w h e n it comes to
survival. We know of -560 supercentenarians
worldwide, and almost 90% are women.
p eople who have lived to be >110 years old
are called supercentenarians. Of the ~560
supercentenarians we know of worldwide, almost
90% are women.
From the 1970s t h r o u g h the 1990s, 3 sites in
the world were reputed to be of special interest
with respect to longevity; these were places
where people were said to have lived unusually
long lives. One was the Caucasus Mountains,
between the Black and Caspian seas at the
southeasternmost part of Europe, where people
were routinely reported to live into their 130s,
140s, 150s, and even 160s. The second was Vil-
cabamba in the Ecuadorian Andes Mountains,
where people did n o t live quite as long...only
into their 130s. The third location was the
Karakoram Mountains, spanning the c o m m o n
borders between India, China, and Pakistan,
80
where the recent earthquake occurred. W h e n
these reports were investigated in detail, n o n e of
t h e m proved true. A telltale clue as to the false-
ness of these claims was that the oldest people
in all of these places were supposedly men, a n d
a n y o n e k n o w i n g a n y t h i n g about h u m a n biolo-
gy knows t h a t is absolutely impossible.
Today, in the United States, Western Europe,
and basically all the technologically developed
countries, w o m e n live longer t h a n men. W h a t
most people do n o t realize is h o w widespread
this difference in longevity is across the coun-
tries of the world. During the time period t h a t
the United Nations has been keeping records,
there have been just 5 countries where m e n
were n o t e d to live longer t h a n w o m e n :
Bangladesh, India, Nepal, Iran, and the Maldive
Islands. W h a t we clearly face in these 5 coun-
tries is a bias in access to nutrition and to health
care. As that bias shrinks, the ratios begin to
reverse (Figure 1). In India, Iran, and the Mal-
dives, w o m e n are n o w living longer t h a n men.
The recent United Nations Demographic Year-
book lists the relative longevity of m e n and
w o m e n in 104 demographic units, ie, mostly
countries but also some islands. 1 Among 103 of
these units in w h i c h life expectancy ranges
from the early 40s to more t h a n 80 years of age,
w o m e n live longer t h a n men. The one c o u n t r y
where m e n live longer? It's Bangladesh.
W h a t a b o u t t h e historical p a t t e r n ? The
answer is t h a t w o m e n have lived longer t h a n
m e n virtually in every place and at every time
we can identify. Sweden has excellent demo-
graphics from the middle of the 18th century,
w h e n life expectancy was in the high 30s
(Figure 2). 2 No matter h o w y o u measure it, no
matter w h e t h e r y o u look at different historical
epochs or at various countries, a very robust
consistent difference is evident. In 1900, life
expectancy in the United States was -48 years,
-46.5 for m e n and -49.0 for w o m e n . Even then,
•/,/aOmen
identify.
have lived longer than men in virtu-
Ily every place and at every time we can
 S.N. Austad

-0- Bangladesh
--O- India
Iran
Nepal
t Maldives
3- Women Live Longer

2-
-'5
E
,,o

0-
¢- A
w
A
U
O-
× -1-
..o
o--

Men Live Longer

-2 i i i i i i i i i i
1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006
Year

Figure 1. Women now live longer than men in a number of countries in which they didn't previously. 1

"0- Absolutedifference
~C~ Percentdifference
14

12
O

I
10
O
E 8

e-"
6

4
X
O
'uO

I I I I I I
1750 1800 1850 1900 1950 2000 2050
'~ ~ Year ~

Life Expectancy (y) 38.4 44.7 71.1 79.7

Figure 2. Women have outlived men in Sweden since records have been kept, beginning in 1751. 2

81
GENDER MEDICINE
there was always a female advantage. It was a
slightly smaller difference as a f u n c t i o n of life
expectancy t h e n compared with now, but the
difference was still there.
Is t h a t difference increasing? The difference
rose from 5% in 1900 to 10% in the year 2000.
But according to US Social Security statistics,
the answer is no. 3 The difference in life ex-
pectancy appears to have peaked a r o u n d 1970
and m i g h t be going down, but it is n o t at all
clear t h a t this is a real trend. For example, if you
compare the sex difference in longevity be-
tween the United States and Japan, the c o u n t r y
with the greatest longevity, you will find that
during the time period w h e n Japan went from
a life expectancy substantially lower t h a n the
United States to substantially higher, Japan had
a gradual widening of the longevity gulf be-
tween m e n and w o m e n (Figure 3). 2 W h e t h e r it
is a short-life country or a long-life country,
w o m e n live longer to variable extents.
I'd like to quickly note m y m e t h o d o l o g y as I
present various kinds of data about the ratio
between male and female longevity, male and
female health statistics, and male and female
O
E
o
u- 5-
4-
'a
c- 3-
o
¢~ 2-
×
ill
..J
0 I I I
1900 1920 1940
Figure 3. Although the sex difference in longevity recently seems to be decreasing in the United States, it has
continued to increase in the world's longest-lived country, Japan. 2
82
mechanistic differences. If y o u have a ratio of 2
quantities, for example, Y1 and Y2, t h a t is a 2-,
3-, 4-, or 5-fold difference, t h e n the way this
appears on a plot depends on w h i c h one y o u
h a p p e n to use as the numerator. If you flip
t h e m and have a 2-, 3-, 4-, or 5-fold difference,
it appears to be m u c h less different simply
because the ratio is n o w b o u n d e d by 0 a n d 1.
I will take the logarithm of those ratios, so it
appears symmetrical. Thus w h e t h e r it is, for
instance, male superiority o n the top a n d
female superiority on the bottom, it does n o t
look different depending on the way I h a p p e n
to express the data.
What Causes the Sex Difference in
Longevity?.
With this in mind, let us ask the question, "Is
this difference in longevity due to one or a few
major diseases, or is it general across a whole
series of diseases?" In casual conversations a lot
of people say, "Well, it has to do with cardiovas-
cular disease." We all know the differences be-
tween the sexes in the development and mani-
festations of cardiovascular disease. However,
United States
--O- Japan
I [ I I
1960 1980 2000 2020
Year

it is m u c h more robust t h a n that. I have taken
6 causes of death and compared t h e m as a func-
tion of age (Figure 4). For virtually all these
causes of death, including accidents, cancer,
heart disease, flu and pneumonia, cerebrovascu-
lar accident, and chronic obstructive p u l m o n a r y
disease, at virtually every age, males die at a
higher rate. These are rather old data, but t h e y
have not changed substantially in terms of the
ratio. This finding holds true for the top 12 caus-
es of death and is not due to just cardiovascular
disease or just cancer or just infectious diseases
or just accidents. 4
or virtually all the most frequent causes of
F death, including accidents, cancer, heart dis-
ease, flu and pneumonia, cerebrovascular acci-
dent, and chronic obstructive pulmonary disease,
at virtually every age, males die at a higher rate.
Is this longevity difference a result of women
aging at a different rate or is there something more
subtle taking place? Women are generally more
0.6- Male Higher Rates
~D
0.4-
O
0
r~
o
O
E 0.2-
~d 0-
0
o, -0.2 -
I I I I
3O 4O 5O
Age (y)
Figure 4. At most ages, males die at a higher rate from numerous causes. 4
S.N. Austad
robust, but they age---that is, they deteriorate---
physiologically at the same rate as m e n do. If you
take the probability that someone is going to die
in the next year as a function of age, the male
and the female difference appears fairly small
(Figure 5A). A person's late 70s and early 80s is
w h e n there is a dramatic increase in this sex dif-
ference. A lot of interesting things are happening
here that are basically masked because I have
expressed the data on a typical arithmetic scale;
therefore, the differences appear small because
the scale of values is huge--there is a 1000-fold
difference between y o u n g and old. If I take the
same data and express t h e m logarithmically, the
differences stand out, particularly those early in
life (Figure 5B).
Birth is a time of relatively high mortality. It
drops d o w n to the pristine peak of physiology
at -10 years of age. After that, it is all downhill,
or on these graphs, it is observed as all uphill.
Note the jump in the late teens and early 20s
that is more p r o n o u n c e d in males. There is a
similar jump in female deaths that is not quite
as dramatic, but it exists. The i m p o r t a n t point,
though, is that this graph--this increase in the
-0- Accidents
--0- Cancer
--A- Heart disease
-~- Flu and pneumonia
-l- Cerebrovascular accident
Chronic obstructive pulmonary disease
Female Higher Rates
I , H ,
6O 7O 80 285
83
GENDERMEDICINE

United States 1999

Male (74.1 }
A --O- Female (79.5)
0.6-

0.5--

0.4
c-

O
0.3
.Z"
,..D 0.2-
_8o
o,_
0.1-

0~ rl!l.l.l~lllf ,.rlWllmlH lllll..UllllUt+ ......

I I I I r I l
0 20 40 60 80 100 120
Age iY)

0.1

O
._Z- 0.01
..,.D

0.001

0.0001 I I I I I I I
0 20 40 60 80 100 120
Age (y)

Figure 5. Sex differences in the probability of dying appear small on an arithmetic scale (A), but can be
clearly seen on a logarithmic scale (B). Numbers in parentheses are mean life expectancies.

probability of death--is one way to t h i n k about
w h a t aging is. It is a way to measure the statisti-
cal sense in w h i c h people deteriorate. If y o u
look at the slope of this increase, y o u can see
that it is virtually identical for males and fe-
males. According to this analysis, females do
n o t live longer t h a n males because t h e y age
more slowly, but rather that females live longer
because t h e y are more robust at every age. If
this difference in the rate of aging does ever

84

begin to disappear, it only occurs very, very late
in life.
emales do not live longer than males because
F they age more slowly,females live longer because
they are more robust at every age.
Historically, if y o u undertake the same kind
of analysis and look at the year 1900, the curves
for m e n and w o m e n are almost on top of one
another. The difference in longevity has to do
with the very slight difference in the elevation,
ie, the starting point, of these curves. It does not
have to do with slope. If y o u take the same kind
of analysis and ask, " W h y do people in some
countries live much, m u c h longer than others?
W h y do Japanese women, for instance, live
-4.5 years longer than American w o m e n ? " It is
again not a difference in the rate of aging, but a
difference in the elevation. The Japanese some-
h o w have managed to avoid such a steep j u m p
at these early ages (15-30 years). W h a t is differ-
ent b e t w e e n m e n and w o m e n in short- and
long-life countries is simply the starting point,
regardless of culture. Apparently, there is a fun-
damental slope that defines h u m a n aging.
he difference in longevity between men and
men in short- and long-life countries is
simply the starting point, regardless of culture;
apparently, there is a fundamental slope that
defines human aging.
The Mortality-Morbidity Paradox
A very interesting paradox a b o u t female
longevity is that despite their lower mortality
rates at virtually every age, w o m e n have higher
overall rates of physical illness, more disability
days, more doctor visits, and more hospital
stays than do men. 4 Almost as interesting as
u n d e r s t a n d i n g the difference in male and
female longevity w o u l d be understanding this
difference in morbidity. Both psychosocial in-
puts and biology are likely to play a role.
S.N. Austad
D espite their lower mortality rates at virtually
every age, women have higher overall rates of
physical illness, more disability dais, more doctor
visits, and more hospital stays than do men.
W o m e n seem to have more severe chronic, but
not life-threatening, illnesses than do men. The
n u m b e r one reason for female disability, particu-
larly later in life at age _>45 years, is arthritis.
W o m e n have higher rates of arthritis throughout
most of what we consider to be midlife and later
life (Figure 6). s The prevalence--the fraction of
the population that has the disease--is n o t the
only issue. There is also a difference in activity
limitations and a difference in hospital visits.
This suggests that w o m e n are not only more
prone to arthritis, b u t also more prone to more
severe arthritis than are men. This is true for a
n u m b e r of chronic conditions, including hear-
ing loss, vision loss, and a variety of n o t life-
threatening b u t nevertheless debilitating condi-
tions. There is a fascinating story to be learned
across these various diseases, b u t at the mo-
ment, we understand very little about w h y these
sex differences exist.
To a biologist like myself w h o is interested
n o t just in people b u t in the rest of the biologi-
cal world as well, an intriguing question is
w h e t h e r female superiority in longevity and
resistance to death b y m a n y individual diseases
is some sort of h u m a n idiosyncrasy or w h e t h e r
it is widely observed in the animal kingdom.
If it is a h u m a n idiosyncrasy, it is one worth
studying, although it w o u l d be a difficult under-
taking because we c a n n o t do m u c h in the w a y
of experiments in h u m a n beings and are limit-
ed to what we can observe. On the other hand,
if this was a general biological p h e n o m e n o n ,
t h e n we could develop animal models and learn
about sex differences in longevity in great detail
b y conducting the appropriate experiments.
Certainly, y o u find the same pattern of sex
differences in longevity in m a n y other species,
for example, the chimpanzee. If y o u analyze
captive chimpanzee demography, females live
longer t h a n males (Figure 7). 6 W h a t about dra-
85
GENDER MEDICINE

- O Men
"-O- Women
700 -
0
600 -
r~
0
0
o 500-
$
a.. 400-
0
300-
c
0

u
200-
c
O
>a 100-
e~

0 I I I I
<45 45-64 65-74 75+
Age Group (y)

Figure 6. After 45 years of age, women have a higher prevalence of arthritis than do men. 5

matically different kinds of species, like tarantu-

las? There y o u have an even more dramatic dif-
ference in longevity. Female tarantulas live rou-
tinely into their 30s, whereas male tarantulas
are relatively short-lived, dying after 5 years.
One of the more spectacular examples is honey-
bees, where males live weeks b u t queens live
years. Apparently, female longevity superiority
is quite widespread in the animal world.
Ilin the animal world, female longevity superiority
quite widespread.
ne hypothesis is that females live longer
O because they have a more active immune
system; however, it is hard to imagine why fe-
males would have a lower incidence of many difi
ferent diseases and attribute differences in all
those diseases purely to their immune system.
male tendency to disperse from their birthplace
to another locale. In captive situations, we hope
to eliminate these mortality risks and examine
differences due to underlying physiology.

nother hypothesis is that estrogen somehow

A key issue is whether we should be concen-
trating on understanding what animals do in
nature, or focusing on what animals do in cap-
tivity. The current belief is that we ought to focus
on what animals do in captivity, because field
studies conflate mortality from physiological
deterioration with other types of mortality. For
instance, in most species of mammals, there is a
great deal of male competition for territories or
for harems. There are a lot of behavioral sources
of mortality, such as male-male combat and the
A protects female animals at a cellular level,
so that females-whether they are rats or mice
or whales or bats or humans-will be generally
longer lived than males.
Hypotheses for Sex Differences in
Longevity
W h y do the species differ? There are 5 mech-
anistic hypotheses about w h y females live long-

86

1.0-(
0.8-
OO
e-
~5
> 0.6-
t,--
0
0.4-
2
0.2-
0 I
10 20
Age (y)
Figure 7. As in humans, chimpanzee females outlive chimpanzee males.6
er that we can begin to evaluate. One hypothe-
sis, which I have heard repeatedly, is t h a t
females live longer because t h e y have a more
active i m m u n e system. The idea is that longevi-
ty is principally determined by i m m u n e func-
tion, and therefore females would be predicted
to live longer. However, it is hard to imagine
w h y females would have a lower incidence of
m a n y different diseases and attribute differ-
ences in all those diseases purely to their
i m m u n e system. A second hypothesis is t h a t
estrogen s o m e h o w protects female animals at a
cellular level, so that females--whether t h e y are
rats or mice or whales or bats or h u m a n s - - w i l l
be generally longer lived.
ccording to the heterogametic sex hypothe-
A sis, the lack of a second X chromosome in
the male will tend to lead to shorter life spans.
The third hypothesis proposes that the het-
erogametic sex, t h a t is, the sex t h a t has 2 differ-
ent sex chromosomes (in mammals, t h a t would
be the XY male) is shorter lived, owing to hav-
S.N. Austad
-O- Female chimpanzees
--O- Male chimpanzees
I I I
30 40 5O
ing only one copy of the larger sex c h r o m o s o m e
(X). In men, for instance, there are 1000+ genes
on the X c h r o m o s o m e that do n o t have an ana-
logue on the Y chromosome. Therefore, a n y
deleterious alleles on the X c h r o m o s o m e will
have no c o m p e n s a t o r y allele in the male,
whereas in females, the second X c h r o m o s o m e
could compensate.
According to the heterogametic sex h y p o t h e -
sis, the lack of a second X c h r o m o s o m e in the
male will tend to lead to shorter life spans. Half
of this hypothesis is based o n this exposed sin-
gle copy of these genes, but the other half is
based on empirical investigation that having 2
X chromosomes m a y be advantageous because
one or the other female X c h r o m o s o m e is in-
activated r a n d o m l y t h r o u g h o u t the tissues, so
there is not an overdose of X-chromosome genes.
Thus, there is a theoretical possibility t h a t as
aging progresses, cells with 1 of the 2 X chromo-
somes t h a t have fewer deleterious alleles--the
better one--will survive better and gradually
predominate in all the cells in a specific tissue.
In other words, because males can o n l y express
the X genes of their single X chromosome,
females could have an advantage by, in princi-
87
GENDERMEDICINE
ple, selecting the better X c h r o m o s o m e while
inactivating t h e deleterious X c h r o m o s o m e . In
fact, there is evidence from peripheral b l o o d in
h u m a n s that the older we are, the more our
cells t e n d to have either the paternal (X P) or
the maternal (X M) c h r o m o s o m e inactivated
rather t h a n having the X c h r o m o s o m e ran-
d o m l y inactivated. That is, the process of X in-
activation gradually changes from being ran-
d o m early in life to b e c o m i n g highly biased
toward either the X M or X P c h r o m o s o m e later
in life. Furthermore, - 1 7 % of the genes on the
"inactivated" X c h r o m o s o m e are n o t fully in-
activated, w h i c h m a y also provide a survival
advantage.
X inactivation gradually changes from being
random early in life to becoming highly
biased toward either the maternal or paternal
chromosome later in life.
The fourth hypothesis, born of a lot of aging
research in recent years, is that animals that
have a reduction in the activity of growth hor-
m o n e and the insulin-like growth factor 1 sig-
naling cascade are longer lived and tend to be
smaller than animals that have greater activity.
This has been studied genetically in detail in
fruit flies, in Caenorhabditis elegans (a round-
worm), and in mice. In each case, w h e n the
activity in this signaling cascade was reduced b y
site-directed mutations along these pathways,
longevity was increased. Anything that is so
general that y o u find it in a 1-mm w o r m as well
as in a m o u s e - - a l t h o u g h it remains to be
observed in humans--strikes me as something
to be taken very seriously.
ne hypothesis,born of recentaging research,
O is that animals that have a reduction in the
actMty of growth hormone and the insulin-like
gro-,vrh factor 1 signaling cascade are longer
lived and tend to be smaller than animals that
have greater activity.
88
e oxidative stress hypothesis of aging pro-
T~p oses that oxygen radicals, produced as an
inevitable consequence of metabolism along with
various defenses against oxidative damage, ul-
timately influence aging and a wide variety of
diseases.
Fifth, the oxidative stress hypothesis of aging,
and one that almost e v e r y b o d y believes has
some validity, is that oxygen radicals, produced
as an inevitable consequence of metabolism
along with various defenses against oxidative
damage, will ultimately influence aging and a
wide variety of diseases. I can tell y o u that in
the aging field, this is the hypothesis that no
one discounts b u t n o b o d y is in love with either.
That is to say, everyone believes that oxidative
stress explains s o m e t h i n g a b o u t aging, b u t few
consider it to be the o n l y m e c h a n i s m of aging.
Evidence For and Against the Hypotheses
To assess the first 2 hypotheses that females
have an e n h a n c e d i m m u n e function or that
estrogen is s o m e h o w generally protective,
species could be examined in w h i c h the sex dif-
ference in life span is more e n h a n c e d than it
is in humans. N o t h o n e y b e e s or tarantulas, be-
cause we do n o t have a great deal of tools to
study those creatures, b u t we could potentially
study other mammals. For example, there are
some dramatic sex differences in longevity
within mammals. In the case of short-finned
pilot whales, females live virtually twice as long
as males. These whales are also interesting be-
cause they are one of the few species to under-
go a true m e n o p a u s e in nature. We are likely to
learn something valuable from studying these
sorts of species that exhibit dramatic differences
in life expectancy.
To assess the heterogametic sex hypothesis,
we can study marsupials, w h i c h unlike the rest
of the mammals, do n o t exhibit r a n d o m X in-
activation. Marsupial females always inactivate
the X P chromosome. Part of the heterogametic
hypothesis involves a gradual selection for one
c h r o m o s o m e versus the other. Marsupials have

lost that advantage. So if this hypothesis was
true, t h e n y o u would expect t h a t there would
n o t be a sex difference in longevity a m o n g mar-
supials or the difference would be smaller t h a n
in the rest of the mammals. Unfortunately, we
do n o t k n o w m u c h about sex differences in
longevity in these animals, with the exception
of a shrew-like marsupial n a m e d Antechinus. In
this species, the males all die w i t h i n 2 weeks
after the end of every breeding season a n d thus
live exactly 1 year. Females, however, live -33%
longer t h a n that. These data in Antechinus are
evidence against the heterogametic sex h y p o t h -
esis, because marsupials should n o t possess dra-
matic female longevity superiority. Males and
females should have very similar life spans.
However, their reproductive biology is very
bizarre, n o t unlike salmon, and very unusual
a m o n g mammals. Their abrupt death is n o t
really aging per se in the sense of the gradual
aging observed in h u m a n s and other animals.
Fortunately, there are examples of some marsu-
pials that do age gradually in the same sense as
we do. For instance, we have good captive data
showing t h a t female Leadbeater's possums do
n o t live longer t h a n the males; in fact, the male
of this species lives -15% longer. Thus, greater
female longevity is not universal. It is n o t even
clear if there are more species in w h i c h females
live longer or vice versa.
reater female longevity is not universal--it is
G even clear if there are more species in
not
which females live longer or vice versa.
Another way to investigate the hypothesis
t h a t the sex possessing the heterogametic chro-
mosomes is going to be longer-lived is to con-
sider birds, because the sex-chromosome situa-
tion is reversed compared with mammals. In
birds, it is the female that has 1 short and I long
sex chromosome, and therefore does n o t have
the backup of the 2 long sex chromosomes (the
Z chromosomes) t h a t the male has. The predic-
tion is t h a t if heterogametic sex is a key factor,
t h e n male birds should be longer-lived. In fact,
S.N. Austad
in 3 species of birds, including budgerigars,
zebra finches, and Japanese quail, males outlive
females, at least in captivity. For every bird
species t h a t I have been able to find in w h i c h
there is good captive data, males outlive the
females. Certainly, this is provocative evidence
that would seem to favor the heterogametic sex
hypothesis. It is of concern, however, t h a t in
some avian species, the female has been report-
ed to outlive the male, but all of these reports
were from field studies and are thus difficult to
interpret for the reasons discussed previously.
n birds, females have 1 short and 1 long sex
I chromosome and therefore do not have the
backup of the males' 2 long chromosomes. In
many species of birds in captivity, males outlive
females-provocative evidence that-would seem
to favor the heterogametic sex hypothesis.
I like the heterogametic sex hypothesis be-
cause it is biologically interesting. Unfortunate-
ly, that does n o t m e a n it is true. There are some
problems with this hypothesis t h a t can be illus-
trated with Brandt's bat, a small bat t h a t weighs
about 7 grams and is a third to a quarter the size
of a mouse. In 1963, a group of researchers in
the Soviet U n i o n w e n t to central Siberia and
marked -1500 of these bats in a particular cave.
Decades later, t h e y went back to the same cave
and f o u n d some of the bats still alive with the
marks on them. The remaining bats t h a t were
still alive were all males up to 41 years of age; no
females were to be found. This is a field study,
however, and as I m e n t i o n e d earlier, we have to
be careful about interpreting field studies. For
instance, does this finding m e a n that the female
bats that were marked just m o v e d somewhere
else? Are t h e y really all dead? Or are the males
alive because t h e y had some sort of different
behavior? Furthermore, males do n o t undergo a
period of extreme food d e m a n d , whereas female
bats are still nursing their y o u n g w h e n their
y o u n g are the same size as they are. We just d o n ' t
k n o w the answers to these questions because we
do n o t k n o w w h a t the underlying physiology is
89
GENDERMEDICINE
and w h e t h e r behavioral differences or physio-
logical differences are responsible for this re-
markable observation in a Siberian cave.
We are also aware of some m a m m a l s in which
the males are significantly longer-lived than the
females; we have very good captive data for 2 of
these species, the guinea pig and the golden
hamster. In b o t h species, the males live substan-
tially longer than the females, thereby contra-
dicting the heterogametic sex and estrogenic
hypotheses. Again, this is a problem in a gener-
al biological sense; it m a y very well be that one
of these h y p o t h e s e s is absolutely valid for
h u m a n s b u t is just n o t generalizable to the rest
of mammals. I would like a general explanation,
and that is something we currently do n o t have.
n both guinea pig and golden hamster, the
I males live substantially longer than the females,
thereby contradicting the heterogametic sex and
estrogenic hypotheses.
Another idea that people have approached
me with is that because w o m e n are smaller than
m e n on average, they live longer. That is, if y o u
control for the sex difference in size and y o u
compare m e n and w o m e n of the same size, the
sex difference in longevity is hypothesized to
disappear. I can tell y o u that the evidentiary
support for this hypothesis is very weak. It m a y
be true, b u t there are just n o t e n o u g h data at
this time to know one way or the other. However,
we do k n o w of species in which males are larg-
er, such as guinea pigs, and species in w h i c h the
females are larger, such as golden hamsters, b u t
in b o t h of these cases, males are longer-lived.
Therefore, the hypothesis that larger animals
Bbothecause we know of species in which males are
larger or in which females are larger, but in
cases the males are longer-lived, the
hypothesis that larger animals have shorter life
spans cannot be a general phenomenon among
mammals.
90
have shorter life spans cannot be a general phe-
n o m e n o n a m o n g mammals.
Lastly, a great deal of mechanistic research
has been c o n d u c t e d on the oxidative stress
hypothesis. The idea is that electrons escape
from the m i t o c h o n d r i a l electron t r a n s p o r t
chain during oxidative p h o s p h o r y l a t i o n and
create oxygen radicals. Extra electrons sticking
to molecular oxygen creates reactive oxygen
species (ROS) such as superoxide radicals.
Antioxidant enzymes like superoxide dismutase
can convert superoxide radicals to h y d r o g e n
peroxide, which, because of Fenton chemistry,
can produce hydroxyl radicals. Hydroxyl radi-
cals are very damaging and are ubiquitous. Be-
cause the creation of damaging oxygen radi-
cals is inescapable and because aging itself is
ubiquitous, the aging c o m m u n i t y believes this
process plays a key role in the whole aging cas-
cade. It is important to note that we n o t only
have the capacity to generate ROS, we also have
the ability to q u e n c h those radicals and repair
their damage. We have, in the mitochondria,
superoxide dismutases that can dismutate su-
peroxide radicals to hydrogen peroxide, while
another antioxidant enzyme, glutathione per-
oxidase, can convert hydrogen peroxide to wa-
ter. But the process of destroying oxygen radi-
cals is n o t perfect. There are some ROS that
escape and cause cellular damage.
The first study that examined sex differences in
oxidative stress production in h u m a n s was re-
search conducted in 2002 in Dr. Barry Halliwell's
lab. 7 Using peripheral b l o o d from a series of
males and females, Halliwell measured oxidative
damage to DNA, because m a n y people believe
that the key site at w h i c h oxidative damage
occurs is in the DNA. Damage occurs in proteins
and lipids as well, b u t DNA is believed to be par-
ticularly critical. W h a t he f o u n d was that the
most c o m m o n oxidative adduct of DNA oc-
curred statistically more frequently in males com-
pared with females. These results are quite pro-
vocative, but observations can only take us so far.
Animal Models of Human Sex Differences
One of the reasons we have n o t made more
progress in mechanisms of sex differences in

longevity is that people have n o t been interest-
ed in studying the question. Just as important,
we also do n o t have a good animal model for
sex difference research in longevity. In mice, the
model that medical researchers love b e y o n d rea-
son, males live longer than females in some
strains, whereas the opposite is true for other
strains. It just depends on the strain. Inbred
strains, of course, are idiosyncratic b y defini-
tion, and that should not be surprising. A m u c h
better model for longevity studies happens to be
rats. However, there are fewer studies with rats
because y o u do n o t have the power of the
genetics that y o u have with mice. Yet, from a
perspective of understanding sex differences,
there is clearly a more consistent sex difference
in longevity in rats. The one study that exam-
ined a variety of mouse genotypes and a variety
of rat genotypes under identical conditions
found male mice from several strains are longer-
lived than the females, and female rats from
several strains are longer-lived than the males
(Figure 8). 8
One of the few labs to undertake sex differ-
ence longevity research is that of Dr. Jose Villa
in Valencia, Spain. He c o n d u c t e d the studies
with Wistar rats, in which he f o u n d that the
15 Mice
O
u
e"
10
0
/ !!
O
~5 -5
-10
-15
Females Live Longer
-20 I I I I
C57BL/6 DBA/2 B6D2F1 B6C3F1
Rodent Genotype
Figure 8. In a study in which 7 genotypes of mice and rats (plus outbred Wistar rats) were treated exactly
alike, male mice lived longer than female mice, but female rats lived longer than male rats. 8
S.N. Austad
females live - 1 6 % longer than the males. This is
near the upper end of the longevity difference
we observe b e t w e e n m e n and w o m e n , and the
kind of difference that we find in countries like
Uzbekistan, Russia, and Kazakhstan. Vifia et al
examined oxidative damage to DNA in isolated
rat liver mitochondria. 9 They f o u n d that male
rats had substantially more DNA damage due to
oxidation than females did. Villa and his col-
leagues also investigated to w h a t extent estro-
gen might be involved in free radical production
as measured by hydrogen peroxide levels. They
observed female rats before and after ovariec-
t o m y and f o u n d that ovariectomy increased the
a m o u n t of oxygen radicals produced. If t h e y
replaced estrogen in the ovariectomized females
with daily subcutaneous injections, they could
actually prevent the increase in h y d r o g e n per-
oxide levels in liver and brain m i t o c h o n d r i a
induced b y ovariectomy. The researchers t h e n
examined e n d o g e n o u s mitochondrial antioxi-
dant levels of glutathione peroxidase. They mea-
sured b o t h mRNA levels and protein activity,
and f o u n d exactly the difference predicted:
antioxidant activity was higher in females and
lower in males. They concluded that estrogen
contributes to the longer life span of the female
Males Live Longer
ii
i
Rats
I I I I I
Wistar F344 BN F344BNF1
91
GENDERMEDICINE
rat because estrogen attenuates the age-induced
increase in ROS production and enhances mech-
anisms of antioxidant protection.
If Vifia and his colleagues had stopped there,
I could leave you with a nice and simple take-
h o m e message. I can't, however, because t h e y
decided to do a quick experiment in mice as
well. Mentioned only in a brief paragraph, Vifia
et al reported that in mice, as in rats, free radi-
cal production was greater in males t h a n in
females. 9 These last results raise serious doubts
about the oxidative stress hypothesis of sex dif-
ferences in longevity, because in mice, males
live longer t h a n females.
Looking Ahead
In conclusion, there is no question that wom-
en outlive m e n in a robust f a s h i o n across cul-
tures, c o n d i t i o n s , a n d causes of death. Female
longevity superiority is c o m m o n in nature, but
certainly n o t ubiquitous. It is n o t even clear, at
least in mammals, whether this sex difference
occurs more frequently in females or in males--
we do n o t have data on e n o u g h species to
answer this fundamental question. Comparative
biology argues against m a n y of the current hy-
potheses explaining w h y w o m e n outlive men.
Even the generally accepted oxidative stress
hypothesis of sex differences in longevity is
hard to support, because although male mice
exhibit greater age-induced oxidative stress
t h a n do females, some male mice outlive female
mice. Currently, the best-supported theory is the
heterogametic sex hypothesis, which explains
the general patterns across species for sex-biased
survival.
We now know which experimental models are
best, and we will soon add a range of new species
to our armamentarium. With the rapid progress
in genome sequencing, Lee Hood has predicted
that within 10 years, a human-sized genome will
Address c o r r e s p o n d e n c e to: Steven N. Austad, PhD, Barshop Institute for Longevity & Aging Studies,
University of Texas Health Science Center, 15355 Lambda Drive, STCBM Bldg, Room 3.100, San Antonio,
TX 78245. E-mail: austad@uthscsa.edu
92
be sequenced in less t h a n a day for about $1000. l°
If this prediction proves true, we will be able to
examine in great mechanistic details some of
the intriguing species in which males live longer
t h a n females or where there is an exaggerated
difference in life span between the sexes. We are
truly on the verge of m a k i n g serious progress in
understanding w h y w o m e n outlive men.
REFERENCES
1. United Nations Statistics Division. Demographic
Yearbook 2003. Available at: http://unstats.un.org/
unsd/pubs/gesgrid.asp?data=year.
2. Universityof California, Berkeley.The Human Mortal-
ity Database. Available at: http://www.mortality.org.
3. Social Security Online Actuarial Publications.
Actuarial studies. October 21, 2005. Available at:
http://www.ssa.gov/OACT/NOTES/actstud.html.
4. Wingard DL. The sex differential in morbidity,
mortality, and lifestyle. Annu Rev Public Health.
1984;5:433-458.
5. Verbrugge LM. Women, men, and osteoarthritis.
Arthritis Care Res. 1995;8:212-220.
6. Hill K, Boesch C, Goodall J, et al. Mortality rates
among wild chimpanzees. JHum Evol. 2001;40:437-
450.
7. Proteggente AR, England TG, Rehman A, et al.
Gender differences in steady-state levels of oxida-
tive damage to DNA in healthy individuals. Free
Radic Res. 2002;36:157-162.
8. Turturro A, Witt WW, Lewis S, et al. Growth curves
and survival characteristics of animals used in the
Biomarkers of Aging Program. J Gerontol A Biol Sci
Med Sci. 1999;54:B492-B501.
9. Borras C, Sastre J, Garcia-Sala D, et al. Mitochon-
dria from females exhibit higher antioxidant gene
expression and lower oxidative damage than males.
Free Radic Biol Med. 2003;34:546-552.
10. Hood L. Systems biology: Integrating technology,
biology, and computation. Mech Ageing Dev. 2003;
124:9-16.