International Review of Psychiatry, October 2010; 22(5): 429–436

Gender differences in depression

GORDON PARKER & HEATHER BROTCHIE

School of Psychiatry, University of New South Wales; and Black Dog Institute, Sydney, Australia

Abstract
It is commonly suggested that a female preponderance in depression is universal and substantial. This review considers that
Int Rev Psychiatry Downloaded from [Link] by National University of Singapore on 06/07/14

proposition and explanatory factors. The view that depression rates are universally higher in women is challenged with
exceptions to the proposition helping clarify candidate explanations. ‘Real’ and artefactual explanations for any such
phenomenon are considered, and the contribution of sex role changes, social factors and biological determinants are
overviewed. While artefactual factors make some contribution, it is concluded that there is a higher order biological factor
(variably determined neuroticism, ‘stress responsiveness’ or ‘limbic system hyperactivity’) that principally contributes to the
gender differentiation in some expressions of both depression and anxiety, and reflects the impact of gonadal steroid changes
at puberty. Rather than conclude that ‘anatomy is destiny’ we favour a diathesis stress model, so accounting for differential
epidemiological findings. Finally, the impact of gender on response to differing antidepressant therapies is considered briefly.

Introduction which have used standardised case-finding measures –

have consistently established much higher rates of
For personal use only.

lifetime depression (both major depression and dys-

Epidemiological studies across a number of thymia) in women (see Weissman et al., 1996;
cultures consistently show that, beginning at Wilhelm, Parke, Geerlings, & Wedgwood, 2008).
puberty, major depression, dysthymia and A representative US study – the National
anxiety disorders are two to three times more Co-morbidity Survey Replication or NCS-R
common in women than in men. (Leibenluft, (Kessler et al., 2005) – which surveyed over 9000
1999) English-speaking respondents, quantified the lifetime
This paper will consider a number of gender risk of ‘any mood disorder’ to be 1.5 times higher for
differences that bear on the epidemiology of depres- women. Thus, the earlier reporting by Weissman and
sion, on the manifestations of depression, on help Klerman (1977) of a female preponderance in depres-
seeking, and on some indicative treatment sion rates has largely been replicated.
differences. There are, however, some intriguing qualifications
to the seemingly consistent finding. Oakley Browne,
Wells, Scott, and McGee (2006) undertook a
Epidemiology regional study of the lifetime prevalence of many
More than thirty years ago, Weissman and Klerman psychiatric disorders in New Zealand. Lifetime rates
(1977) wrote a definitive epidemiological review were significantly higher for women than for men for
demonstrating a distinctly higher prevalence of major depression (20.3% versus 11.4%) and for
depression in women in community and clinical dysthymia (2.6% versus 1.6%), but with the overall
samples, and which they judged as reflecting both gender ratio for ‘any mood disorder’ (24.3% versus
real and artefactual factors. As women are more 15.6%) being somewhat attenuated as a consequence
likely to seek help for depression, such a factor would of a slightly lower rate of bipolar disorder in women
artefactually influence the female preponderance in (3.6% versus 4.1%). Several studies have suggested
clinical samples, but would largely be circumvented that rates of bipolar I disorder are comparable across
in community samples – and argue (by default) for the genders (e.g. Kessing, 2004) but that rates of
the difference principally being contributed to by bipolar II disorder are somewhat higher in women
‘real’ factors. (e.g. Benazzi, 1999). When identified, such a gender
Large community studies subsequently conducted difference in bipolar disorder is less distinctive than
over the last three decades in differing countries – and described for major depression.

Correspondence: Gordon Parker, School of Psychiatry, University of New South Wales; and Black Dog Institute, Sydney, Australia. Tel: 61293824372.
Fax: 61293824343. E-mail: [Link]@[Link]
ISSN 0954–0261 print/ISSN 1369–1627 online ß 2010 Institute of Psychiatry
DOI: 10.3109/09540261.2010.492391
 430 G. Parker & H. Brotchie
While ‘melancholia’ is variably viewed as a sepa- interesting review (Stevenson & Wolfers, 2009) of
rate depressive sub-type or a ‘more severe’ expres- community survey studies, not only were women less
sion of clinical depression, and is quite variably ‘happy’ than men, but the data indicated that the
defined on the basis of its clinical features, some difference has been widening over recent decades.
commentators (e.g. Kessing, 2005) have suggested In representative studies, individuals were asked
that there is no clear evidence of gender differences ‘In general, how satisfied would you say you
in the prevalence of melancholia. Thus, any female personally are with your life today?’ Study findings
preponderance in clinical depressive conditions is indicated that there had been ‘a pervasive decline in
less obvious or non-existent in the quintessential the relative well-being of women in nearly all
biological disorders (i.e. melancholia and bipolar European countries’. For the USA the authors
I disorder). quantified that ‘a man in 2006 is happier than a
Further, when socially homogeneous (in terms of similarly situated man was 34 years ago, while
education, culture, parity, etc.) population sub-sets women. . .are less happy by a roughly equal
Int Rev Psychiatry Downloaded from [Link] by National University of Singapore on 06/07/14

are surveyed, gender differences are not always amount’. The review noted that such findings were
observed – with illustrative samples including evident across women who remained at home as well
British civil servants (Jenkins, 1985), the Amish as women who were in employment, and judged that
community (Egeland & Hostettet, 1983), college the ‘common thread appears to be an increasing
students (Grant et al., 2002) and a UK Jewish ambition of young women beyond the domestic
community (Loewenthal et al., 1995). sphere’. They suggested that women were under
The latter finding allows some important caveats – increasing pressure to be ‘successful’ at multiple
firstly, that while ‘clinical depression’ (as a domain) levels, including at work, as a family member, in
appears more common in women, the female making a contribution to society, and in being a
preponderance may not be consistent across all community leader. Clearly, such findings would
depressive sub-types, and that it is more evident in weight social stress factors, alone or in activating
the unipolar non-melancholic depressive disorders any predispositional diathesis factors.
For personal use only.

(which we conceptualize as reflecting the impact of Pursuit of causal factors is assisted by examining
precipitating life events on predisposing personality for age effects. In community studies where gender
or temperament styles). Even within the differences are identified, such differences are not
non-melancholic groups, it might be expected that constant across age groups. In pre-adolescents, rates
there would be variable gender ratios – and later we of depression are low and either comparable for boys
will consider ‘atypical depression’, where the female or girls, or slightly higher rates are reported in boys
preponderance appears particularly distinctive. The than in girls (Jorm, 1987a). Any such excess in young
second caveat, that gender differences are most boys might reflect a real difference or be an artefact
evident in heterogeneous community samples and of reporting or observation. As a generalization,
may be absent in samples of socially homogeneous when young girls become depressed they tend to go
groups, again argues for the relevance of social quiet and keep to themselves, while boys are more
factors, whether proximal and/or distal. Such epide- likely to act out with anger and irritability, with such
miological findings have encouraged us to hypothe- ‘externalizing’ behaviours in boys resulting in their
size (as detailed later) that a ‘diathesis stress’ model ‘depression’ being more likely to be observed by
may operate. In essence, rather than argue that others and so artificially inflating the rate of
‘anatomy is destiny’, and that women are, of neces- ‘observed depression’ in boys. Jorm (1987a) further
sity, more likely to develop depression, we suggest quantified males and females as having their highest
that women are merely more predisposed (i.e. the rates in their twenties (at 17% and 26% respectively)
‘diathesis’) and that social factors (i.e. the ‘stress’ and falling at a steady rate from 40 years onwards.
component) modulate that predisposition. In socially Most epidemiological studies that have examined
heterogeneous communities we assume that women the impact of age on gender differences have indicated
experience more precipitating social factors and/or that the female preponderance commences at around
are more vulnerable to certain social factors a 10–12 age band but that such a differentiation is far
(so accounting for findings in large community more distinctive for ‘major depression’ than it is for
studies), whereas in socially homogeneous commu- ‘dysthymia’ (Parker & Hadzi-Pavlovic, 2004).
nities, gender-specific social precipitants are less Following that separation emerging at puberty, the
relevant or have less impact. peak differentiation would appear to be from 15 to
There is one other domain that is rarely considered 30 years of age. From the late thirties onwards, rates
in reviews of the general proposition that women are for both men and women decrease progressively,
more likely to be depressed than men. In the last although the gender difference persists and never fully
decade there has been increasing scientific interest in reaches parity. Such studies of lifetime rates
the topic of ‘happiness’ or ‘well-being’. In an predictably rate initial and recurrent episodes
 Gender differences in depression 431
or ‘prevalence’. When rates of first onset or ‘incident’ that over time men were more likely to ‘forget’
depression are quantified, as undertaken by Parker episodes (absolutely, or ‘remember’ fewer symptoms
and Hadzi-Pavlovic (2004) in an analysis from the US so that episodes no longer met ‘case’ criteria), while
National Co-morbidity Study or NCS (Kessler et al., women were more likely to ‘remember’ episodes that
1994), a similar pattern in the gender difference had generally not been nominated or had not
emerging in adolescence is evident. In addition – but previously reached ‘case’ criteria. Such a difference
in women only – there was a second distinct increase may reflect differential rates of denial or other
in rates of initial episodes of major depression in their reporting biases that have been described in men,
late forties and early fifties. Such a secondary peak is as well as women being particularly more likely to
compatible with the historical description of ‘involu- show ‘mood amplification’, with Briscoe (1982)
tional melancholia’, in that it was previously suggested reporting that women were more likely to report
that women (in particular) were likely to develop an higher levels of both negative and positive affects.
episode of depression around menopause. As the Measurement influences may also contribute. For
Int Rev Psychiatry Downloaded from [Link] by National University of Singapore on 06/07/14

bimodal incidence peaks in depression rates in example, depression measures that have items such
women correspond to expected ages of menarche as crying (which is much more likely to be experi-
and menopause, sex role hormones and/or psychoso- enced and reported by women), or appetite and
cial factors are suggested. weight changes (when women are more likely to
Such rich epidemiological findings allow interpre- engage in comfort eating), will inflate depression
tative attempts a firmer footing, and serve to shape rates. Similarly, and as reviewed previously by us
the next section. (Wilhelm & Parker, 1989), women are more likely to
rate as ‘cases’ if the criterion is the actual number of
symptoms, with Angst and Dobler-Mikola (1984)
How do the epidemiological findings
therefore arguing for the need to convert the current
inform us?
weighting given to symptom numbers in diagnostic
Such epidemiological findings are capable of two measures. In our cohort study (Wilhelm, Parker, &
For personal use only.

broad explanations – in essence, that gender differ- Asghari, 1998), we quantified that women reported
ences are (to some extent) artefactual or essentially more symptoms than men and were more likely to
real. The artefactual possibility essentially supposes report certain symptoms (e.g. feeling self-critical,
that, as women are more likely to seek help, volunteer irritability, tearfulness, experiencing appetite gain
depressive features when specifically asked, report and weight changes).
differentially when seeking help and respond differ- Angst and Dobler-Mikola (1984) demonstrated
entially to depression rating measures, they will score that the distinct preponderance of depression in
as more depressed purely as a consequence of such women was attenuated when social impairment cri-
artefactual influences. As noted, however, commu- teria were deleted, again suggesting that much of the
nity surveys (where a female preponderance is contribution to the female preponderance relates to a
consistent) are unlikely to involve a ‘help seeking’ gender difference in experiencing and/or reporting
component. This artefactual factor can therefore be affect and mood symptoms, and weighting symptom
effectively discounted as having explanatory power in severity as against social and work impairment.
community survey populations. However, the possi- In conclusion, there are a number of artefactual
bility that women might, in clinical contexts where factors (help-seeking, mood amplification, differen-
help-seeking components may be operative, be more tial experiential and symptom influences on presen-
likely to ‘volunteer’ depressive symptoms must be tation, and measurement nuances) that are likely to
conceded. As we will later note that gender differ- contribute to the quantified higher rates of ‘depres-
ences in depression only relate to first episode and sion’ in women.
not to repeat episodes, the suggestion of a
‘help-seeking’ explanation appears less viable as a
Beyond artefactual factors, what might be
substantive explanation.
‘real’ determinants of gender differences?
We have previously described one intriguing
gender difference in reporting. In pursuing the The emergence of a female preponderance in clinical
relevance of artefactual influences (Wilhelm & depression in late childhood or early adolescence
Parker, 1994) in our longstanding cohort study, we could reflect (1) sex role changes, (2) social factors,
used a case-finding measure to quantify lifetime and/or (3) biological factors.
episodes of depression at 5-year intervals and with The ‘sex role’ hypothesis is that certain gender-
lifetime depression rates of cohort members being based role experiences shape the development of
assessed and re-assessed at each review. Such a ‘self’ and, as a consequence, any diathesis to anxiety
strategy allowed us to check on their consistency of and depression. For example, pink baby clothes for
reporting episodes over time. In essence, we found girls and blue clothes for boys, providing dolls for
 432 G. Parker & H. Brotchie
girls and war figures for boys with which to play and not merely onset. However, the gender differ-
might lead to girls having a greater propensity to ence is restricted to incidence or initial onset. This
‘internalize’ or be more passive and less ‘mentally might allow support for a diathesis stress model –
tough’ than boys, and therefore create a stronger with an emphasis on the predisposing diathesis
predisposition to anxiety or depression. factor, rather than on any precipitating social factor
The ‘social factor’ hypothesis argues that women components. Our previously referenced cohort study
are either differentially exposed to a greater number is therefore worthy of detailing and consideration.
of life event stresses and/or are more vulnerable to In 1978, we recruited a sample of university
them. One exemplar is the generalization that students (all having completed a primary degree
women preferentially invest their self-esteem in an and engaged in their final year of teacher training),
intimate dyadic relationship whereas men have more and we have now followed them for some thirty
diffuse self-esteem investments – being spread across years. In an earlier paper (Wilhelm et al., 1997) we
intimate and social relationships as well as invested in reported on their 15-year review, when the sample
Int Rev Psychiatry Downloaded from [Link] by National University of Singapore on 06/07/14

their work. The hypothesis therefore supposes that, had a mean age of 39 years. While there was a trend
by women putting all their eggs (here self-esteem) in for a female preponderance in lifetime rates of ‘major
one basket, they are more vulnerable to depression if depression’ and ‘all depression’, neither trend was
their investment is put in jeopardy (e.g. their significantly different. As they had all trained as
husband neglects, abuses or leaves them). This teachers (with most continuing in that profession)
hypothesis is indirectly supported by gender differ- and were socially homogeneous in terms of social
ences in depression being influenced by marital class, we again interpreted the failure of a gender
status (Tennant, Bebbington, & Hurry, 1982), with difference to emerge over time as reflecting their
married men having lower rates of depression then social homogeneity in terms of less exposure to
single and divorced men while married women are gender-specific stressors that may be more likely to
more likely to have the highest rates of those three be experienced by women in a general community.
groups, and has led to the aphorism that ‘marriage is We therefore argued for a diathesis stress model
For personal use only.

toxic to women and protective to men’. for the gender difference. In essence, that women
While such sex role and social factor hypotheses have a greater predispositional vulnerability to
have some intuitive appeal and with the latter depression and to certain salient social factors that
(in particular) likely to make some contribution precipitate or activate a depressed state, but that in
(as earlier overviewed), there are also strong argu- socially homogeneous groups that diathesis might
ments against them being distinctive contributors. If not be effected. A candidate predispositional per-
the sex role hypothesis was valid we should anticipate sonality factor is ‘neuroticism’, as women tend to
a gender difference to emerge at a young age (i.e. in score relatively higher on this personality/tempera-
childhood) and not, as occurs, at adolescence. ment domain and it is recognized as contributing to
Secondly, in relation to social factors, Kessler depression and anxiety (Wilhelm & Parker, 1989).
(2000) showed that after controlling for rape and Unfortunately, the term ‘neuroticism’ has negative
other sexual traumas in the NCS survey, the gender and ad hominem connotations, and in considering its
difference was halved, while the gender differential contribution later, we will use alternative terms such
was restored by broadening social stressor variables as ‘limbic system hyperactivity’.
to include traumatic experiences more likely to be Turning to biological explanations, we should first
experienced by men. Thus, certain social factors may consider whether it is likely that there is (1) some
have quite differing salience to men and to women factor that specifically disposes to a gender difference
and, depending on the prevalence of such gender- in depression or (2) a higher order factor that
specific factors (e.g. bullying, sexual abuse, rape, loss disposes to depression and also to some other
of employment, work harassment) in any commu- (circumscribed) conditions. If the latter model is
nity, might contribute to the gender difference in valid, then not only is the question widened but the
prevalence. search for explanations is narrowed.
However, there is a key argument for rejecting the Many years ago we contemplated whether there
possibility that sex role or social factor hypotheses were other conditions that showed a similar epide-
make a distinctive contribution. Kessler (2000) miological pattern. Some physical conditions
examined survey data from the NCS and other (e.g. migraine) showed some pattern similarities,
community studies and showed that the male while most psychiatric conditions either showed a
preponderance in lifetime depression rate was only quite contrasting male preponderance (e.g. socio-
for the first depressive episode and not for second pathy, drug and alcohol dependency) or no gender
and recurrent episodes. If gender differences were difference. However, certain anxiety disorders
underpinned by sex role or social factors we would showed a distinctly parallel epidemiological pattern,
expect gender to impact on both onset recurrence, compatible with the model postulating that the
 Gender differences in depression 433
gender difference in depression might more reflect a the nature of ‘neuroticism’? Again, in our review
gender difference in a ‘higher order’ diathesis vari- (Parker & Brotchie, 2004), we noted quite
able which then sequentially impacted ‘downstream’ wide-ranging descriptions, including a neurotic char-
on depression and some anxiety conditions. Some acter style, autonomic lability, a down-regulated
overview of that model testing hypothesis is now HPA axis, brain activation response to negative
provided. stimuli (in neuroimaging studies) and ‘stress respon-
As Breslau, Schultz, and Peterson (1995) judged siveness’ (across sympathetic and parasympathetic
pre-existing anxiety disorder as a ‘potential factor’ in autonomic system measures, and ‘harm avoidance’.
the emergence of gender differences in major depres- Consolidating these descriptions, a construct of
sion, we (Parker & Hadzi-Pavlovic, 2001) accessed ‘emotional responsiveness’ was clearly suggested.
the US community survey NCS database (involving We then overviewed hormonal changes at puberty,
nearly 6,000 subjects) and examined the extent to both in females and in males, examining for pubertal
which gender and ‘any prior anxiety disorder’ hormonal changes that lead to greater ‘limbic system
Int Rev Psychiatry Downloaded from [Link] by National University of Singapore on 06/07/14

influenced the risk of subsequent depression. hyperactivation’ (i.e. the diathesis factor) in women,
Survival analyses indicated that ‘prior anxiety’ made so conferring greater responsivity to negative emo-
a stronger contribution than ‘gender’, both in relation tional stimuli, and leading to a greater likelihood of
to major depression (risk of 1.4 and 1.2 respectively) certain anxiety and depressive states. That review
and to dysthymia (risk of 1.5 and 1.2). Such analyses considered the possible contributions of oestrogen
both implicated the impact of anxiety on subsequent and progesterone, and their impact on several
depression and, in indicating that the influence of primary neurotransmitter systems (i.e. the locus
prior anxiety on depression was similar for both males ceruleus-norepinephrine system, the serotonin
and for females, suggested a very parsimonious system, and the GABA-benzodiazepine receptor
diathesis factor. In a second analysis of the same complex), and referenced studies demonstrating
NCS database, we (Parker & Hadzi-Pavlovic, 2004)
that brain sites associated with anxiety (e.g. limbic
examined the age of onset (by gender) for those who
For personal use only.

forebrain) have binding sites for gonadal steroids. We

received a range of differing anxiety and depressive
therefore argued that post-pubertal women are more
disorder diagnoses. While some anxiety conditions
likely to demonstrate limbic system hyperactivation
(i.e. agoraphobia, social phobia, simple phobia and
in response to negative stressors as a consequence of
panic attacks) showed a trend for a female prepon-
gonadal hormone influences, with higher ‘neuroti-
derance emerging before puberty, two anxiety disor-
cism’ scores being a marker of the propensity to
ders (i.e. generalized anxiety and panic disorder)
emotional responsiveness, and that such a diathesis
showed a distinct increase (and with the increase even
generates the differences in prevalence rates of
more distinctive in girls than boys) in early adoles-
cence and which corresponded temporally with the anxiety and depression.
emergence of a female preponderance in depressive Others have articulated a similar model. For
disorders (particularly major depression). We there- example, Young and Korszun (1998) noted that
fore argued that there may be a higher order common ‘Organisational differences between male and female
factor (and specifically a gender hormone-based one) brains are caused by exposure to gonadal steroids in
that influenced the differential expression of both the pre- and perinatal periods. The interaction of
certain expressions of anxiety and depression at the these organisational effects in females with cyclical
time of puberty. steroid hormone changes after puberty, followed
We (Parker & Brotchie, 2004) therefore over- then by menopause and the loss of these same
viewed candidate temperament styles, with neuroti- steroids, suggest that stress responsiveness and sus-
cism, as noted earlier, emerging as a clear candidate. ceptibility to stress-related disorders could vary
Jorm (1987b) reported a meta-analysis of studies substantially over women’s lifetimes’. They went on
examining gender differences in neuroticism scores, to state that ‘it is possible. . .that the increased stress
and quantified that, while females had higher scores responsiveness of females, contributes to the
across all age groups, differences were more distinc- increased prevalence of anxiety disorders and auto-
tive in young and middle-aged adults than in nomic hyper-arousal in women compared to men.’
children and the very old. The curvilinear pattern We suggested (Parker & Brotchie, 2004) that such
for neuroticism scores (by gender) was very similar to a gender-based difference might, from an ethological
that reported by Jorm (1987a) for depression. We perspective, have certain advantages. Such a beha-
then overviewed data indicating that neuroticism was vioural pattern of vigilant defence would assist the
(1) a strong risk factor for depression, and for first individual’s survival and the survival of those under
episodes of depression, (2) the strongest predictor of their care. Such a propensity would be of distinct
depression persistence in primary care, and (3) advantage to women (in their role as mothers) but a
modifiable by social factors. In this context, what is hindrance to the males engaged in requisite ‘hunter
 434 G. Parker & H. Brotchie
tribesman’ independent activities, and where fear- (1) anxious, (2) irritated, (3) stressed, suggesting in
lessness would have group survival advantages. essence that there was settling of ‘emotional dysre-
gulation’. We then examined the personality profiles
of those who reported chocolate craving when
Manifestations of depression
depressed. Importantly, there was no relationship
Here we expand on some material presented earlier to personality or temperament styles that emerged
in relation to any artefactual impact on measurement from a higher order introversion style. However, all
in community surveys. personality styles that emerged from a higher order
A number of studies (e.g. Simpson, Nee, & ‘neuroticism’ style were associated with reporting of
Endicott, 1997) have examined for gender differ- chocolate craving when depressed, and again we
ences in depressive patterns, with most studies implicated a personality style of ‘limbic cortex
showing no distinctive differences or only the occa- hyperactivity’ or ‘emotional dysregulation’ as both
sional difference. In the latter studies, reports have contributing to depression, the aforementioned
Int Rev Psychiatry Downloaded from [Link] by National University of Singapore on 06/07/14

variably implicated women as more likely to experi- homeostatic coping repertoire and to a consequential
ence hypersomnia, hyperphagia and weight gain, impact on the gender difference in depression
feelings of guilt and worthlessness, fatigue ability, prevalence.
tearfulness, tension and somatic pain.
In terms of coping repertoires in response to
Treatment differences
depression, the general finding is that women are
more likely to internalize (e.g. go quiet, go to their We have recently reviewed (Parker et al., in press)
bedroom, cry) and men are more likely to externalize literature examining whether there are gender differ-
(e.g. show anger, increase their alcohol intake). Such ences in response to differing psychotherapies for
differences may reflect coping per se and/or the depression. We found no consistent trend for women
impact of antecedent ‘comorbidities’). In relation to or men to report greater responsiveness to differing
the last point, in a report from the STAR*D study, psychotherapies or counselling in the literature. In
For personal use only.

Marcus et al. (2005) quantified that women reported our own web-based survey (Parker & Crawford,
more anxiety and somatoform disorders, bulimia and 2009), we did find that, while males did not report
atypical symptoms, while alcohol and drug abuse any antidepressant treatment modality as preferen-
were more common in men. tially beneficial, women were more likely to rate two
Across the differing depressive sub-types, gender of the three nominated psychological therapies (i.e.
differences in manifestations are most evident in CBT and counselling) as more effective. If a valid
‘atypical depression’. In this condition (which is difference, this could suggest that women may be
defined in DSM-IV (APA, 1994)), individuals are more likely to reflect their greater affiliative tenden-
held to have a highly reactive (or responsive) mood, cies and capacity to form a treatment alliance more
be more likely to experience hyperphagia and readily in therapy, as against men being more
hypersomnia, be more likely to experience a sense defensive and guarded.
of ‘leaden paralysis’ and have a personality style However, even if valid, any such propensity might
marked by ‘sensitivity to rejection’. We have previ- theoretically be modified by the gender of that
ously argued (Parker et al., 2002) that the personality therapist. In pursuing the latter possibility, we
style is the primary feature, and that the atypical reported (Parker & Hyett, 2009) two studies. In the
features of hypersomnia and hyperphagia may more first study we merely analysed referral letters to the
be homeostatic features rather than symptoms. Black Dog Institute’s depression clinic and quanti-
The homeostatic interpretation was argued on the fied female practitioners as writing more detailed
basis that the hyperphagia is weighted to certain referral letters than male practitioners (i.e. 220 versus
foods (such as carbohydrates and chocolate) which 88 words). In the second study we surveyed 500
release endorphins (promoting ‘feel good’ sensa- individuals who had experienced an episode of
tions), and other gut and brain peptides, while the clinical depression in the previous year and sought
hypersomnia restores slow wave sleep during stress. assistance from their general practitioner. In this
Pursuing the nature of ‘atypical depression’, we study we quantified that female general practitioners
undertook a study (Parker & Crawford, 2007) where were perceived as more caring then male general
some 3000 individuals who had experienced an practitioners and more likely to refer the patient to a
episode of clinical depression, completed a web- mental health professional. By contrast, male general
based survey. When depressed, 54% reported food practitioners were more likely to show poor eye
cravings with chocolate craving being far more contact, to cut off the patient, to be verbally distant
distinctive in women (51%) than in men (31%). and to suggest a solution before hearing the patient
For those who craved chocolate, they were distinctly out. Further, they were more likely to prescribe a
more likely to report that it made them feel less drug as monotherapy and their female patients were
 Gender differences in depression 435
less likely to wish to return to see them than to a depression rates. The topic would benefit from
female general practitioner. In essence, this study further explanation of the hypothesis that there may
suggested that there were also gender differences in be a higher order impact (rather than being limited to
managing depression, and that women might be depression per se) and the capacity of differing
preferentially more sensitive to those factors treatments to modify the higher order diathesis
than men. factor.
Returning to differential treatment responses,
there has been the suggestion (Kornstein et al., Declaration of interest: This paper was funded by
2000) that women may be more likely to respond to an National Health and Medical Research Council
selective serotonin reuptake inhibitors (SSRIs) than a programme grant (510135) and a New South Wales
tricyclic antidepressants, and with the converse Department of Health Infrastructure Fund. The
phenomenon holding for men. The differential authors report no conflicts of interest impacting on
gender response to SSRI antidepressants was again this topic, and they alone are responsible for prepar-
Int Rev Psychiatry Downloaded from [Link] by National University of Singapore on 06/07/14

suggested in an analysis from the STAR*D study ing this paper.

(Young et al., 2009), but has not been confirmed in a
number of other studies as reviewed by Thase et al.
(2000), and would benefit from refined studies (see References
also Keers & Aitchison, 2010 this issue).
American Psychiatric Association (1994). Diagnostic and statis-
tical manual of mental disorders (DSM-IV) (4th ed).
Washington, DC: American Psychiatric Association.
Conclusions
Angst, J., & Dobler-Mikola, A. (1984). The definition of depres-
The gender difference in unipolar depression rates sion. Journal of Psychiatric Research, 18, 401–406.
Benazzi, F. (1999). Gender differences in bipolar II and unipolar
has been a relatively (but not absolutely) consistent
depressed outpatients: A 557-case study. Annals of Clinical
finding for over three decades and led to a rich set of Psychiatry, 11, 55–59.
studies that have sought to contemplate determi-
For personal use only.

Breslau, N., Schultz, L., & Peterson, E. (1995). Sex differences in

nants. This review seeks to suggest that the phe- depression: A role for preexisting anxiety. Psychiatry Research,
nomenon is somewhat dependent on the type of 58, 1–12.
depression, that the gender difference is not Briscoe, M.E. (1982). Sex differences in psychological well-being.
Psychological Medicine Monograph, 1, 1–46.
restricted to depression (being mirrored in data for Egeland, J.A., & Hostettet, A.M. (1983). Amish Study I. Affective
‘neuroticism’ and for some anxiety disorders), that disorders within the Amish, 1976–1980. American Journal of
quantified data can be influenced by a range of Psychiatry, 140, 56–61.
artefactual factors, and that even after conceding Grant, K., Marsh, P., Syniar, G., Williams, M., Addlesperger, E.,
social factor contributions there would appear to be Kinzler, M.H. et al. (2002). Gender differences in rates of
depression among undergraduates: measurement matters.
strong biological underpinnings. Further, the rich- Journal of Adolescence, 25, 613–617.
ness of studies across community surveys and Jenkins, R. (1985). Sex differences in minor psychiatric morbidity.
homogeneous social groups argues for a diathesis Psychological Medicine, 7, S1–53.
stress model, and with a strong contribution from Jorm, A.F. (1987a). Sex and age differences in depression:
gonadal hormones contributing to the diathesis A quantitative synthesis of published research. Australian and
New Zealand Journal of Psychiatry, 21, 46–53.
factor of ‘emotional dysregulation’ being Jorm, A.F. (1987b). Sex differences in neuroticism: A quantitative
over-represented in women (and emerging at pub- synthesis of published research. Australian and New Zealand
erty), and disposing women to a greater propensity to Journal of Psychiatry, 21, 501–506.
both anxiety and depression. Keers, R, & Aitchison, K. (2010). Gender differences in antide-
Women are not, of necessity, more likely to pressant response. International Review of Psychiatry, 22(5),
485–500.
develop depression than men. There would appear Kessing, L.V. (2004). Gender differences in the phenomenology
to be a biological diathesis factor that predisposes of bipolar disorder. Bipolar Disorders, 6, 421–425.
women from puberty to be at greater risk but this Kessing, L.V. (2005). Gender differences in patients presenting
diathesis is modified by socio-cultural factors. with a single depressive episode according to ICD-10. Social
Preliminary data indicating that women and men Psychiatry and Psychiatric Epidemiology, 40, 197–201.
Kessler, R.C. (2000). Gender differences in major depression:
may differ in response to differing drug and non-drug
Epidemiological findings. In E. Frank (Ed.), Gender and its
therapies argue for more substantive studies to effects on psychopathology (pp. 61–84). Washington DC:
investigate this intriguing possibility and implications American Psychiatric Association.
for managing the depressive disorders. Kessler, R.C., Berglund, P., Demler, O., Jin, R.,
Merikangus, K.R., & Walters, E.E. (2005). Lifetime prevalence
and age-of-onset distributions of DSM-IV disorders in the
Future directions National Comorbidity Survey Replication. Archives of General
Psychiatry, 62, 593–602.
There have been sufficient epidemiological studies Kessler, R.C., McGonagle, K.A., Zhao, S., Nelson, C.B.,
considering and quantifying the impact of gender on Hughes, M., Eshleman, S. et al. (1994). Lifetime and 12-
 436 G. Parker & H. Brotchie
month prevalence of DSM-III-R psychiatric disorders in the Parker, G.B., & Brotchie, H.L. (2004). From diathesis to
United States. Archives of General Psychiatry, 51, 8–19. dimorphism: The biology of gender differences in depression.
Kornstein, S., Schatzberg, A.F., Thase, M.E., Yonkers, K.A., The Journal of Nervous and Mental Disease, 193, 210–216.
McCullough, J.P., Keitner, G.I. et al. (2000). Gender differ- Simpson, H.B., Nee, J.C., & Endicott, J. (1997). First-episode
ences in treatment response to sertraline versus imipramine in major depression: Few sex differences in course. Archives of
chronic depression. The American Journal of Psychiatry, 157, General Psychiatry, 54, 633–639.
1445–1452. Stevenson, B., & Wolfers, J. (2009). The paradox of declining
Leibenluft, E. (1999). Gender Differences in Mood and Anxiety female happiness. American Economic Journal: Economic Policy,
Disorders: From Bench to Bedside. Washington DC: American 2, 190–225.
Psychiatric Association. Tennant, C., Bebbington, P., & Hurry, J. (1982). Female
Loewenthal, K., Goldblatt, V., Gorton, T., Lubitsch, G., vulnerability to neurosis: The influence of social roles.
Bicknell, H., Fellows, D. et al. (1995). Gender and depression Australian and New Zealand Journal of Psychiatry, 16, 135–140.
in Anglo-Jewry. Psychological Medicine, 25, 1051–1063. Thase, M.E., Frank, E., Korstein, S., & Yonkers, K.A. (2000).
Marcus, S.M., Young, E.A., Kerber, K.B., Kornstein, S., Gender differences in response to treatments of depression. In
Farabaugh, A.H., Mitchell, J. et al. (2005). Gender differences E. Frank (Ed.), Gender and its effects on Psychopathology
Int Rev Psychiatry Downloaded from [Link] by National University of Singapore on 06/07/14

in depression: Findings from the STAR*D Study. Journal of (pp. 103–129). Washington, DC: American Psychiatric Press.
Affective Disorders, 87, 141–150. Weissman, M.M., & Klerman, G.L. (1977). Sex differences and
Oakley Browne, M.A., Wells, J.E., Scott, K.M., & McGee, M.A. the epidemiology of depression. Archives of General Psychiatry,
(2006). Lifetime prevalence and projected lifetime risk of 34, 98–111.
DSM-IV disorders in Te Rau Hinengaro: The New Zealand Weissman, M.M., Bland, R.C., Canino, G.J., Faravelli, C.,
Mental Health Survey. Australian and New Zealand Journal of Greenwald, S., Hwu, H.-G. et al. (1996). Cross-national
Psychiatry, 40, 865–874. epidemiology of major depression and bipolar disorder.
Parker, G., & Crawford, J. (2007). Chocolate craving when Journal of the American Medical Association, 276, 293–299.
depressed: A personality marker. British Journal of Psychiatry, Wilhelm, K., & Parker, G. (1989). Is sex necessarily a risk factor to
191, 351–352. depression? Psychological Medicine, 19, 401–413.
Parker, G., & Crawford, J. (2009). Personality and self-reported Wilhelm, K., & Parker, G. (1994). Sex differences in lifetime
treatment effectiveness in depression. Australian and depression rates: Fact or artefact? Psychological Medicine, 24,
New Zealand Journal of Psychiatry, 43, 518–525. 97–111.
Parker, G., Crawford, J., & Blanch, B. (2010). Does gender Wilhelm, K., Parker, G., & Asghari, A. (1998). Sex differences in
influence response to differing psychotherapies by those with the experience of depressed mood state over fifteen years.
For personal use only.

unipolar depression? Journal of Affective Disorders (in press). Social Psychiatry and Psychiatric Epidemiology, 33, 16–20.
Parker, G., & Hadzi-Pavlovic, D. (2001). Is any female Wilhelm, K., Parker, G., & Hadzi-Pavlovic, D. (1997).
preponderance in depression secondary to a primary female Fifteen years on: evolving ideas in researching sex differences
preponderance in anxiety disorders? Acta Psychiatrica in depression. Psychological Medicine, 27, 875–883.
Scandinavica, 103, 252–256. Wilhelm, K., Parker, G., Geerligs, L., & Wedgwood, L. (2008).
Parker, G., & Hadzi-Pavlovic, D. (2004). Is the female pre- Women and depression: A 30-year learning curve. Australian
ponderance in major depression secondary to a gender and New Zealand Journal of Psychiatry, 42, 3–12.
difference in specific anxiety disorders? Psychological Medicine, Young, E.A., & Korszun, A. (1998). Psychoneuroendocrinology
34, 461–470. of depression. Hypothalamic-pituitary-gonadal axis. Psychiatric
Parker, G., & Hyett, M. (2009). Management of depression by Clinics of North America, 21, 309–323.
general practitioners: Impact of physician gender. Australian Young, E.A., Kornstein, S.G., Marcus, S.M., Harvey, A.T.,
and New Zealand Journal of Psychiatry, 43, 355–359. Warden, D., Wisniewski, S.R. et al. (2009). Sex differences
Parker, G., Roy, K., Mitchell, P., Wilhelm, K., Mahli, G., & in response to citalopram: A STAR*D report. Journal of
Hadzi-Pavlovic, D. (2002). Atypical depression: A reappraisal. Psychiatric Research, 4, 503–511.
The American Journal of Psychiatry, 159, 1470–1479.