Depression remains a leading cause of distress and disability worldwide. In the 1997 Australian National Survey of Health and Wellbeing, 7.2% of people surveyed had experienced a mood (affective) disorder in the previous 12 months.1 Those affected reported a mean of 11.7 disability days (when they were “completely unable to carry out or had to cut down on their usual activities owing to their health”) in the previous 4 weeks. There was also evidence of substantial undertreatment: only 35% of people with a mental health problem had a mental health consultation during the previous 12 months. Of those with a mental health problem, 27% (ie, three-quarters of those seeking help) saw a general practitioner. In the 2007–08 Australian National Health Survey, not much had changed: 12-month prevalence rates were 4.1% for depression, 1.3% for dysthymia and 1.8% for bipolar disorder.2 These disorders were associated with significant disability, role impairment, and mental health and substance use comorbidity. Again, there was evidence of substantial unmet need, and again GPs were the health professionals most likely to be providing care. However, one change in the intervening period was a reduction in treatment with antidepressants.

While GPs have many skills in the assessment and treatment of depression, they are often faced with people with depression who simply do not get better, despite the use of proven therapies, be they psychological or pharmacological. This supplement aims to address some of the issues that GPs face in this context.

To set the scene, Shadow Minister for Finance Andrew Robb (doi: 10.5694/mjao12.10777) has provided an “insider’s view” of depression. We are grateful to Mr Robb for his eloquent exposition of what it actually feels like to have a depressive illness. His account should serve as an ongoing reminder that depression happens to a wide cross-section of people, that those people suffer, and that our therapeutic endeavours should always be delivered in the context of the individual. GPs are well placed in this regard, as they often have a longitudinal knowledge of the patient, understand his or her circumstances, stressors and supports, and can marshal this knowledge into a coherent and comprehensive management plan. Of course, GPs should not soldier on alone if they feel the patient is not getting better. The rest of this supplement outlines strategies that they might adopt in such cases.

In trying to understand what happens when GPs feel “stuck” while treating someone with depression, Jones and colleagues (doi: 10.5694/mjao12.10566) report a qualitative study that aimed to gauge the response of GPs to the term “difficult-to-treat depression”. They found that, while there was confusion around the exact meaning of the term, GPs could relate to it as broadly encompassing a range of individuals and presentations. Thus, the term has face validity, if not specificity. More specific terms such as “treatment-resistant depression” are generally reserved for a subgroup of people with difficult-to-treat depression that has failed to respond to treatment, with particular management implications.

A set of articles then tackles one scenario in which depression can be difficult to treat — in the context of physical illness. Depression is often expressed via physical symptoms, but the obverse is that people with chronic physical ailments are at high risk of depression, as outlined by Olver and Hopwood (doi: 10.5694/mjao12.10597). Pain syndromes, as discussed by Holmes and colleagues (doi: 10.5694/mjao12.10589), are particularly tricky, as complaints of pain require the clinician to accept them as “legitimate”, even if there is no obvious physical cause. The use of analgesics can create its own problems, including dependence. Patients with comorbid chronic pain and depression require careful and sensitive management and a long-term commitment from the GP to ensure consistency of care and support.

Depression (and anxiety) can also be seen in the context of cancer. This is in many ways understandable, given the existential issues raised by a cancer diagnosis, along with the pain and discomfort associated with the cancer and, often, its treatment. Couper and colleagues (doi: 10.5694/mjao12.10522) provide useful strategies to assist in assessing mood in people with cancer (eg, how to distinguish depression from demoralisation), as well as articulating treatment options.

Interest in the role of lifestyle factors in the development and persistence of mental illnesses is gaining momentum. Jacka and Berk (doi: 10.5694/mjao12.10508) detail the literature that supports an important role for exercise and diet in the management of depression, and provide useful practical tips to assist GPs in educating their patients about these issues. This should be an integral part of the comprehensive care of people with depression and can also help overcome associated physical health issues, such as obesity.

A series of articles then discusses the topic of psychiatric comorbidity in depression. In clinical practice, this is more common than not, and underappreciation of such issues can lead to suboptimal treatment outcomes. Beatson and Rao (doi: 10.5694/mjao12.10474) tackle the difficult topic of depression co-occurring with borderline personality disorder (BPD). People with BPD have, as part of the core disorder, a perturbation of affect associated with marked variability of mood. This can be very difficult for the patient to deal with, and can feed self-injurious and other harmful behaviour. The key to management is the provision of a therapeutic “holding environment” and therapeutic consistency. Use of mentalisation-based techniques is gaining support, and psychological treatments such as dialectical behaviour therapy form the cornerstone of care. Use of medications tends to be secondary, and prescription needs to be judicious and carefully targeted at particular symptoms. GPs can play a very important role in helping people with BPD, but should not “go it alone”, instead ensuring sufficient support for themselves as well as the patient.

Tiller (doi: 10.5694/mjao12.10628) provides an overview of anxiety in the context of depression. This is a very common comorbidity, can be associated with poor clinical outcomes and carries particular therapeutic challenges. A careful balancing of psychological and pharmacological interventions is required for optimal success. Another particularly problematic form of depression is that which occurs in the context of bipolar disorder. As Berk and colleagues (doi: 10.5694/mjao12.10611) show, firm data on how best to manage bipolar depression are surprisingly lacking. It is clear that treatments such as unopposed antidepressants can make matters a lot worse, with the potential for induction of mania and mood cycle acceleration. However, certain medications (notably, some mood stabilisers and atypical antipsychotics) can alleviate much of the suffering associated with bipolar depression. Specialist psychiatric input is often required to achieve the best pharmacological approach. For people with bipolar disorder, psychological techniques and long-term planning can help prevent relapse. Family education and support is also an important consideration.

Schizophrenia is an often disabling condition associated with a broad lack of integration in society, low levels of income, lack of gainful employment, and high rates of physical health problems (notably cardiovascular disease) and substance use. It is hardly surprising that people with schizophrenia are at high risk of depression, but this is often missed clinically or the symptoms are simply attributed to the core illness. Bosanac and Castle (doi: 10.5694/mjao12.10516) provide clinical clues to assessing mood in people with schizophrenia, and outline potential therapeutic strategies. GPs play a crucial role in the comprehensive care of people with schizophrenia, and a high index of suspicion for depression should be part of this.

Curran and colleagues (doi: 10.5694/mjao12.10567) address what is likely to become a more common scenario with an ageing population — depression in the context of dementia. This is a complex clinical presentation, and requires vigilance, careful monitoring, and appropriate liaison with other members of the care team and the patient’s family. The treatment of depression in people with dementia requires particular care, and attention must be given to avoiding any deleterious effects of prescribed antidepressants or other agents.

The final three articles in this supplement cover treatment strategies for patients with depression that does not respond adequately to usual first-line and second-line treatments. Chan and colleagues (doi: 10.5694/mjao12.10495) provide a succinct account of potential pharmacological interventions for treatment-resistant depression, which has failed to respond to an adequate course of an antidepressant, although they point out the paucity of definitive evidence-based research in this area. This is a scenario in which GPs should consider specialist psychiatric review for their patients. Fitzgerald (doi: 10.5694/mjao12.10509) details non-pharmacological biological approaches to difficult-to-treat depression. Electroconvulsive therapy has a long history in this regard and still holds an important place in the management of severe depression. However, the emerging technology of magnetic seizure therapy shows promise as an alternative application with fewer cognitive side effects. Other emerging treatments, such as repetitive transcranial magnetic stimulation and transcranial direct current stimulation, have the potential to alleviate depression in some people, with relatively few side effects. For a very select group of patients with depression, vagus nerve stimulation and deep brain stimulation have provisional evidence of effectiveness.

Psychological therapies such as cognitive behaviour therapy and interpersonal therapy have strong support in research and clinical practice as effective treatments for depression, either alone or in conjunction with antidepressant medications. Casey and colleagues (doi: 10.5694/mjao12.10629) show how such therapies can be adapted for use in people with severe depression, and also outline broader psychosocial aspects of care.

The range of articles in this supplement highlight that a broad and individually focused plan of care, tailored to the person’s needs and clinical profile, is the cornerstone of quality care. We thank all contributing authors and hope that this supplement assists clinicians generally, and GPs in particular, to enhance the health outcomes of patients with difficult-to-treat depression.

It is two years since I started enjoying the best mornings I have had since childhood, after finally finding a treatment for my underlying depressive condition.

I was 12 years old when I first recognised that something was hampering my ability to function properly early in the day. I would wake up feeling very negative, lacking confidence, indecisive, and just wanting to stay within myself. But I would find that, after having been up and about for a few hours, my state of mind would dramatically improve. I would become positive, a decisionmaker, happy to take responsibility.

For years, I simply put my malaise down to “not being a good morning person”, but as I got older, the fog would take longer to clear. It started to have an impact well into my working day.

I didn’t want to admit, even to myself, that I had a depressive condition. Rather than seeking professional help, I did my utmost to self-manage my condition, to try to get myself going. I employed a range of quite bizarre techniques, including staring at the sun to make myself sneeze, because I felt that would help with the release of endorphins and give me a lift.

I also have the biggest collection of positive-thinking books that you will ever see.

I found I felt better when the adrenaline was pumping, and I think this is why I have always subconsciously been attracted to vocations with a fair share of drama and crises. Business and politics both fit that bill nicely.

Looking at the sun, or my other strategy of putting a pen in my mouth to simulate a smile, was not normal behaviour of course. But back then, I was concerned by the great stigma that was attached to depression, or to any mental health condition for that matter. There still is stigma, but it is diminishing.

Mental health issues were always seen as a sign of character weakness or fragility, particularly in men. Yet I was at my best in a crisis, when the adrenaline kicked in.

It took me 43 years to confront my condition. Initially, I went to see a local general practitioner. I was told not to bother taking pills, as they didn’t really work — it’s all in the head! I left the consultation feeling deflated and thinking, “get over it, it’s something I’ll just have to live with”.

After my condition continued to worsen, I decided to pick up the phone to a friend, Jeff Kennett, who has done so much good work with beyondblue. On his advice, I saw a highly respected mental health specialist, who soon discovered that I release serotonin, a hormone that helps kickstart your day, about four or five hours later than the average person. For some people, as with me, this period of chemical imbalance can be associated with a depressive condition.

It took six months of experimentation with different types of medication, but I finally found a treatment that worked. I now combine this with daily exercise. I’m one of those questionable people who jump in the bay at dawn to swim 1200 to 1500 metres, even in the middle of winter. There is no better way to get the heart started.

Once I was confident that I had overcome my problem, I agreed to write a book giving my account of living life with a depressive condition. The book, Black dog daze, is really a series of anecdotes based on different memories and events throughout my life, from childhood and time on the farm where I grew up, through to my time working for the Packer family, and then entering politics.1

I made the decision to write the book after my specialist advised me that, during a regular get-together of mental health professionals, several of them reported a spike in the number of middle-aged men seeking diagnosis and treatment after my battle with depression became public.

I am determined to take any opportunity to spread the word that help is available, and there is no good reason to do what I did and what I suspect thousands of others do — cover it up and suffer in silence for decades. On the other hand, my experience also confirms that it is never too late to confront your condition.

You don’t have to tell the world — just go along quietly and get some professional advice, and consult those you love and trust. There is light at the end of the tunnel. Your life can be transformed.

Depressive symptoms frequently accompany physical illness, but the association between the two is complex. Medically ill patients with depression have reductions in quality of life,1,2 increased medical morbidity and mortality,3 increased functional disability,4,5 reduced occupational performance,4 and reductions in role functioning. Other implications of depressive comorbidity include prolonged hospital admission, amplification of physical symptoms,6,7 reduction in adherence to medical treatment,8 and increased medical costs and health care use.9,10 Here, we focus on recognition, diagnosis and management of comorbid depression in patients with medical illness.

It has been estimated that rates of depression rise from 2%–5% in the community to 5%–10% in primary care settings and 6%–14% in hospital inpatients.11 Reported rates of depression vary considerably between individual medical illnesses (Box 1). High rates of depression have been noted in patients with neurological disorders, particularly epilepsy (20%–55%), multiple sclerosis (40%–60%) and stroke (14%–19%). There is also considerable variability of reported rates of depression within individual medical illnesses. For example, reported rates of depression in patients with idiopathic Parkinson disease (IPD) vary from 2.7% to 90%.17 Demonstrating more realistic estimates, a systematic review found that 17% of patients with IPD had major depression, 22% had minor depression and 13% had dysthymia.18

Depression in physically ill patients presents challenges in recognition and diagnosis. One reason for this is the difficulty in accounting for physical symptoms. Depressive symptoms are often clustered into biological (somatic), psychological and social symptoms (Box 2). Many diagnostic tools and scales for depression rely heavily on the presence of somatic symptoms, including insomnia, loss of appetite, anergia and reduced libido, to confirm a diagnosis. However, the somatic symptoms associated with depression are frequently also present in many physical illnesses, causing uncertainty about their attribution to either depression or the physical illness. In a similar way, social symptoms of depression, including withdrawal and impairments in role functioning, may occur in both depression and physical illness, making them unreliable as markers of depression in medically ill patients. The diagnosis of depression in medically ill patients therefore relies more heavily on symptoms of psychological distress, including preoccupation with guilty themes and failure, impairments in self-esteem, and an inability to derive joy from previously enjoyed activities.19

A second issue reducing recognition of depression in physically ill patients is the uncertainty surrounding the extent of symptoms. Physical illness is inevitably associated with losses, and depressive symptoms such as sadness, worry and irritability may be judged by both patients and primary carers as an “understandable” reaction to these losses, and subsequently minimised in importance. This kind of misjudgement by the clinician can lead to either underdiagnosis of depression or overdiagnosis and inappropriate prescribing of antidepressants, adding to the side-effect load of other medications.

Several depression rating scales have been used as screening tools to enhance the recognition of depression in medically ill patients. The Center for Epidemiologic Studies Depression Scale (CES-D) is a 20-item self-report measure that takes around 5 minutes to complete and includes only four somatic items. Good sensitivity and specificity in medically ill populations have been reported with cut-off scores of 16–17;20 however, low positive predictive value (PPV) limits its use as a screening instrument for depression. The Hospital Anxiety and Depression Scale (HADS) is a 14-item self-report scale from which an anxiety score and a depression score can be derived. It was developed specifically for measuring depressive symptomatology in medically ill populations.21 The depression subscale is based on self-reported anhedonia (the inability to derive enjoyment). The HADS has similar reported psychometric properties to the CES-D and also takes only 5 minutes to complete. The latent structure of the HADS has recently been questioned, and it may be best understood as a reliable general measure of distress in medically ill patients.22 The nine-item self-report Patient History Questionnaire (PHQ-9) was originally developed as the initial screening tool of the PRIME-MD (Primary Care Evaluation of Mental Disorders), a diagnostic instrument for anxiety, depression and eating disorders in primary care.23 The PHQ-9 has been trialled in large populations of medically ill patients, with good sensitivity, specificity and PPV. The Beck Depression Inventory (BDI) has been extensively used as a screening tool for depression, but it is lengthy (taking 10–15 minutes to complete) and includes a number of somatic items, making it less helpful for use in medically ill patients.24

Other very brief screening tools for depression in medically ill populations have been developed, including “the distress thermometer” (see Couper et al, ),25,26 single-item questions such as “Are you depressed?” or “Do you often feel sad or depressed?”, and the two-item version (PHQ-2) of the PHQ-9 that asks the patient for a self-report regarding loss of interest or pleasure and feeling “down, depressed or hopeless”. The brief single-item tools are limited by reports of lower sensitivity, specificity and PPV, although the PHQ-2 shows promise. Unfortunately, few of the abovementioned screening tools (apart from the CES-D) have been validated against the most important standard, which is a psychiatric interview and associated clinical diagnosis.

Depressive symptoms are frequently encountered in patients with physical illness. This comorbidity has widespread implications and presents some management problems. Recognition of depression in medically ill patients is the first barrier to treatment, and an emphasis should be placed on psychological symptoms when making the diagnosis. Once recognised, several interventions for depression have been shown to be effective in this patient group, including minimising the disability of the medical illness, general supportive measures, individual CBT, and antidepressant medications in patients with moderate to severe symptoms. When using antidepressant medications, it is important to consider drug–drug and illness–drug interactions when choosing the class of medication and the starting dose.

Borderline personality disorder (BPD) is a serious mental illness characterised by dysregulation of emotions and impulses, an unstable and inconsistent sense of self and of others in close relationships, and marked difficulties in interpersonal relationships, often accompanied by suicidal and self-harming behaviour.

The instability of emotions and predominance of negative affect that characterise BPD often lead to problems determining whether the patient has major depression co-occurring with BPD, or whether the depressive symptoms are part and parcel of the BPD itself. In this article, we aim to assist clinicians facing this situation to make an accurate diagnosis.

To prepare the article, we searched PsychInfo and MEDLINE databases for articles published between 2000 and 2012 relating to BPD co-occurring with major depression, other depressive disorders, or bipolar disorder. Review articles and those involving randomised controlled trials of treatment were particularly sought. Book chapters relevant to the search criteria were also examined.

Major depressive disorder (MDD) commonly co-occurs with BPD. The lifetime prevalence of major depression in the course of BPD was 83% in one large study,1 which accords with other research and clinical experience.

Patients with BPD often present to clinicians with depressive symptoms. As the symptoms of depression and BPD overlap significantly, it can be challenging to make an accurate diagnosis of a major depressive illness when the disorders co-occur. Accurate diagnosis is essential because each disorder requires treatment in its own right. It is important to note that rating scales of depression, whether patient- or clinician-rated, are less helpful for assessing the severity of depressive symptoms when BPD is present.2

Given the prominent overlap of symptoms between BPD and affective disorders, it has been suggested that BPD is a variant of affective disorder — either depressive disorder or bipolar disorder.3,4 However, the consensus of expert opinion is that BPD is not a variant of either MDD or bipolar disorder,5–12 although overlap of symptoms of both disorders can occur with BPD.

The most significant evidence that BPD is not a variant of depressive disorder is that treatment of depression does not result in remission of BPD symptoms. An important longitudinal study found that effective treatment of BPD tends to result in remission of depression, and antidepressants often show only modest benefit for depressive disorders that co-occur with BPD.6

A 2010 review of phenotype, endophenotype and genotype comparisons between BPD and MDD found that BPD differs from MDD in symptomatology, prognosis, heritability, patterns of brain region involvement, neurohormonal indices and sleep architecture.7 Some biological processes overlapped, including amygdala hyperreactivity, volume changes in the anterior cingulate cortex, and deficient serotonin function. The authors noted that definitive clarification of the commonalities and differences between BPD and MDD requires examination of both disorders using the same study design and methodology.

A study of depressive symptoms and BPD features in dysthymic disorder showed that a common factor underlying both disorders best explained the frequency of their co-occurrence, providing an excellent fit with the data.13 The authors postulated that the underlying factors were shared genetic, temperamental, or early environmental risk factors.

It is known that factors in the early environment, including those that lead to insecure and pathological patterns of attachment, combined in some cases with an anxious, sensitive temperament and later childhood trauma, predispose to both BPD and early-onset dysthymic disorder and depression.14 It may be that some of these factors also predispose to rapid-cycling bipolar disorder.

Family studies show that, while MDD and bipolar disorders commonly co-occur with BPD, impulsive spectrum disorders are more common than affective spectrum disorders in BPD-affected families.10,11 Studies have also shown that depressed patients with BPD do not significantly differ from those without BPD in terms of morbid risk for bipolar disorder in first-degree relatives.10,11 Importantly, treatment with mood-regulating agents does not produce remission from BPD.10

A recent review examining the similarities and differences between bipolar affective disorder and BPD concluded that they are separate entities but have significant elements in common.12 The many differences between them included those relating to sense of self, relationship disruptions, family history of bipolar disorders, benefits of medication, and the form of affective dysregulation.

Another recent review of the overlap between bipolar disorder and BPD found the greatest overlap occurred in relation to rapid-cycling bipolar disorder.15 Frequent shifts of mood characterised both BPD and rapid-cycling bipolar disorder, as did a history of childhood trauma including sexual, physical and emotional abuse, and neglect. Maladaptive self-schemas of the “I am bad” type were much more characteristic of BPD, as was vulnerability to abandonment. The review’s authors question whether rapid-cycling bipolar disorder is a form of BPD or represents bipolar disorder with co-occurring BPD traits.

Clearly, there are unanswered questions about the reasons for the frequent co-occurrence of affective disorders and BPD, which can only be resolved by further research.

Depressive symptoms that occur as part of BPD are usually transient and related to interpersonal stress (eg, after an event arousing feelings of rejection). Such “depression” usually lifts dramatically when the relationship is restored. Depressive symptoms in BPD may also serve to express feelings (eg, anger, frustration, hatred, helplessness, powerlessness, disappointment) that the patient is not able to express in more adaptive ways. The patient’s “depression” in these cases represents a maladaptive endeavour to communicate his or her unhappiness about a particular person or situation. Such depressive states will not respond to antidepressant treatment, but to careful elucidation of the underlying feelings, followed by assisting the patient to address the problem in more adaptive ways.

On cross-sectional assessment, the transient depressive symptoms of BPD may be indistinguishable from symptoms of a major depressive episode (MDE). This can lead to incorrect diagnosis in the absence of a longitudinal history.

A longitudinal history, with careful examination of the depressive symptoms over recent days and weeks, is required to make an accurate diagnosis of MDE or MDD co-occurring with BPD. The diagnostic criteria for MDE in patients with BPD are the same as those for MDE in the general population: consistently lowered mood for at least 2 weeks, significant impairment of energy, lowered interest in usual activities, sleep disturbance (which may involve insomnia or hypersomnia), weight loss or gain, increased suicidal ideation (sometimes accompanied by increased suicidal or other self-harming behaviour), anhedonia, sense of worthlessness, poor concentration and attention, and impairment of functioning (social, occupational or other).16 Melancholic and psychotic features are seldom seen in MDE co-occurring with BPD. However, although the overall pattern of symptoms is the same as in the general population, the quality of the depression in BPD is different.17–20 Depression in BPD is characterised by the features shown in Box 1.

It should be stressed that a diagnosis of MDE should be made in a patient with BPD only when the patient’s mood has been consistently lowered for a minimum of 2 weeks, in conjunction with other symptoms sufficient to meet the criteria for MDE.

We are aware of no research specifically examining medication for major depression co-occurring with BPD. The consensus of informed opinion over many years has been that depression co-occurring with BPD does not respond as well to antidepressant medication as depression in the absence of BPD.5,21 In addition, there is evidence that the time to remission tends to be longer and recurrence of depression more likely when BPD and MDD co-occur.22 A meta-analysis of the outcome of depression co-occurring with personality disorders in general similarly concluded that: “Combined depression and personality disorder is associated with a poorer outcome than depression alone”.23 The percentage of patients with BPD in the sample included in this analysis was not stated.

However, not all authorities agree that depression co-occurring with BPD responds poorly to antidepressant treatment.24,25 A 2002 review found that whether or not personality disorder worsened treatment outcome depended on the study design — the best-designed studies reported the least effect of personality disorder on the outcome of treatment for depression.25 Individual personality disorders were not examined separately. High neuroticism scores were found to be predictive of poor prognosis, particularly when long-term outcome was taken into account. High neuroticism scores are characteristic of BPD,26 and relapse of depression tends to be earlier and time of remission shorter in BPD,22 suggesting that patients with BPD and co-occurring depression may fall in the group identified in this review as responding poorly to treatment for depression.

We agree that vigorous treatment of depression is required when it co-occurs with BPD,25 to ensure the best possible outcome for the patient, but believe that this must be combined with treatment for the co-occurring BPD. In the absence of adequate data, clinicians should consider treating MDD associated with BPD with biological treatments (antidepressants), as they would treat MDD without BPD. However, without BPD-specific psychotherapy, MDD that is associated with BPD may not respond adequately to biological treatments — but BPD-specific psychotherapy does help treat both MDD and BPD when the disorders co-occur.

There is some limited evidence for the use of aripiprazole, olanzapine and omega-3 fatty acids in the management of depressive symptoms of BPD,21 but there are no data to guide clinicians in choosing a specific biological treatment for MDD that co-occurs with BPD. Lithium has not been shown to be particularly effective in treating MDD that co-occurs with BPD. Clinicians may need to exercise caution in prescribing lithium for patients with BPD and MDD, given the significant risk (about 25%) of self-harm through overdose of prescription medication in patients with BPD.27

Rigorous research examining the treatment response of patients with BPD and co-occurring depression is clearly required and should include patients with severe forms of BPD, who are likely to be among those responding most poorly to treatment for co-occurring depression.

There is increasing pressure worldwide to limit the use of medication for BPD because of its limited effectiveness and concerns about the obesity-related health problems that can occur, particularly with polypharmacy.21 Guidelines published by the United Kingdom’s National Institute for Health and Clinical Excellence in 2010 go so far as to state that drug treatment should not be used for individual symptoms or behaviour associated with BPD,28 whereas the latest Cochrane review states that pharmacotherapy should be targeted at specific symptoms only.29 Both guidelines acknowledge the possible benefit of medication when depression and BPD co-occur, while stressing that a diagnosis of depression should be reviewed when it has been made in the context of a crisis in the patient’s life.

Unfortunately, polypharmacy is commonly seen in patients with BPD, with or without co-occurring depression. It tends to occur because the patient’s distress is of such intensity that clinicians feel driven to try to alleviate it, by whatever means are available. Such means often include increased doses of medication or additional medications. The ensuing danger is that patients with BPD may be prescribed one psychotropic agent after another, sometimes in high doses, with none of the earlier prescriptions ceased. One study found that, compared with people with major depression alone, people with BPD were twice as likely to have received anti-anxiety medication, more than six times as likely to have received mood stabilisers, more than 10 times as likely to have used antipsychotics, and twice as likely to have taken antidepressants.30

All available evidence points to the fact that medication should be targeted at specific symptoms of BPD and used only when these are severe, and that medications should be reviewed on a regular basis.21,28,29

There is little evidence that electroconvulsive therapy (ECT) is beneficial for treating depression in patients with BPD.21 There are no randomised controlled trials of its use, and studies have in general been poorly designed. A 2004 study employing improved methodology continued to show a poorer acute response to ECT for depression co-occurring with BPD.31 However, if patients with BPD have severe depression that would otherwise warrant ECT, they should not be excluded from consideration of this intervention simply on the basis of having BPD.

The principal treatment for BPD is psychosocial — that is, some form of psychotherapy, which may be combined with psychotropic medication aimed at specific symptoms.32,33

One study showed that MDD is not a significant predictor of outcome for BPD, but BPD is a significant predictor of outcome for MDD.6 This finding led the study authors to recommend that treatment of BPD with psychotherapy should take priority when BPD co-occurs with depression, as once the BPD abates, so will the depression.6

Several modalities of psychotherapy for BPD, including dialectical behaviour therapy, mentalisation-based treatment, transference-focused therapy, schema-focused therapy and supportive psychotherapy, have been shown in randomised controlled trials to result in lowered levels of depression.33 It is unclear whether the depression referred to in these studies is part of an MDE co-occurring with the BPD or the depressive symptoms so often seen in BPD. These psychotherapies share some common features that are applicable across all treatment settings where patients with BPD are likely to present, including primary care (Box 2). There is no doubt that interactions with patients with BPD that lack these core features will worsen their distress and can lead to increasingly maladaptive (including self-harming) behaviour.

It is significant that those psychotherapies that have proven useful for the treatment of depression alone — cognitive behaviour therapy, mindfulness-based approaches, and interpersonal therapy, among others — all focus on promoting the patient’s capacity to reflect on his or her own mind and that of others (ie, to “mentalise”).34 In some cases, a lowered rate of relapse of depression after psychotherapy has been shown. This lower relapse rate is likely to result from the patient’s new-found ability to reflect on the thoughts, feelings and behaviour associated with, or leading to, depression.34

We suggest taking a therapeutic stance to MDD co-occurring with BPD that promotes reflection on the patient’s mind and the minds of others, using a psychological approach such as cognitive behaviour therapy, mindfulness, interpersonal psychotherapy or mentalisation-based treatment. Such reflection can, over time, limit the repetition of the maladaptive ways of thinking, interacting and behaving that lead to depression and depressive symptoms in patients with BPD. This approach, combined with the features common to effective therapies for BPD shown in Box 2, will help patients to understand themselves and others better, and can lead to lasting change, including less depression.

Medication for MDE, when it co-occurs with BPD, should be prescribed with careful attention to the dangers of polypharmacy and in the knowledge that medication is not the primary treatment for BPD. Leading authorities stress that medication for depression co-occurring with BPD will not lead to remission of the personality disorder,6,21 which requires treatment in its own right.

Depression and anxiety disorders are among the most common illnesses in the community and in primary care. Patients with depression often have features of anxiety disorders, and those with anxiety disorders commonly also have depression. Both disorders may occur together, meeting criteria for both. It can be difficult to discriminate between them but it is important to identify and treat both illnesses, as they are associated with significant morbidity and mortality. General practitioners are well placed to identify and take a primary role in treatment of these illnesses, to facilitate better mental health outcomes.

In Australia, the 12-month prevalence of anxiety disorders is 14.4% and of affective disorders, 6.2%.1 It has been demonstrated that 39% of individuals with generalised anxiety disorder (GAD) also meet criteria for depression.2 About 85% of patients with depression also experience significant symptoms of anxiety, while comorbid depression occurs in up to 90% of patients with anxiety disorders.3 Considering anxiety and depression symptoms together in a “mixed anxiety-depressive disorder” has been proposed,4 but it is useful to identify and institute effective treatment for each set of symptoms. They can occur in all age groups — almost 50% of older adults with 12-month history of GAD met criteria for lifetime major depressive disorder, while only 7.4% of those without GAD met these criteria.5 Both anxiety and depression are associated with substance use disorder,6 and about 7% of the affected population represent serious cases with high comorbidity.7

Up to 25% of the patients seen in general practice have comorbid anxiety and depression.8 Though recognised in both rural and non-rural primary care, there is often a treatment gap, with patients undertreated for either or both disorders.9 Patients with anxiety and/or depression are particularly likely to present with physical complaints rather than mental health symptoms,10 and symptomatology may initially seem vague and non-specific.

Developmentally, anxiety disorders are almost always the primary condition, with onset usually occurring in childhood or adolescence.11 Comorbidity of anxiety and depression is explained mostly by a shared genetic vulnerability to both disorders, or by one disorder being an epiphenomenon of the other.12

Increased corticotropin-releasing factor in cerebrospinal fluid has been reported in both anxiety and depression, but other peptides or hormones of the hypothalamic–pituitary–adrenal axis are regulated differently in the two disorders.13 More recently, neuroinflammatory, oxidative and nitrosative pathways have been implicated in depression and its comorbidities.14 It is most likely that the first episode of depression in a person’s life follows a psychosocial stressor. After three or more episodes, it becomes increasingly likely that subsequent episodes are spontaneous rather than following an external event.15

Comorbid depression and anxiety can increase impairment16,17 and health care use,18 compared with either disorder alone. Their co-occurrence is often associated with a poor prognosis19 and significant detrimental impact on functioning in the workplace.20 The number and severity of anxiety symptoms, rather than the specific anxiety diagnosis, correlate strongly with the persistence of subsequent depressive symptoms, and this relationship is stable over decades (Box 1).21

Many people do not seek treatment for anxiety and depression and, when they do, treatments are not always used effectively. Australian data suggest that 40% of people with current disorders did not seek treatment in the previous year and, of those who did, only 45% were offered a treatment that could be beneficial.22 Despite the high prevalence of depression and anxiety, and notwithstanding Australia’s universal health insurance scheme, service utilisation in this country is low. An Australian survey published in 2001 indicated that only 35% of people with a mental disorder had consulted a health professional for a mental health problem during the previous year, but most had seen a GP for that disorder or for some other health reason.23 In this setting, barriers to effective care were stated to be patient knowledge and physician competence.

It is important to delineate the specific depressive disorder and the specific anxiety disorder, as each may require different interventions. The most prevalent anxiety disorders in Australia are post-traumatic stress disorder (6.4%), social phobia (social anxiety disorder; 4.7%), agoraphobia (2.8%), GAD (2.7%), panic disorder (2.6%) and obsessive–compulsive disorder (1.9%). Of the population aged 16–85 years, 14.4% have an anxiety disorder. The prevalence of depression is 6.2%, with the prevalence of unipolar depressive episodes being 4.1%, dysthymia, 1.3%, and bipolar disorder, 1.8%.1 Some patients have two or more disorders.

Diagnostic criteria are designed to distinguish between disorders, and exclude clinical features that are common to more than one. Thus, criteria for depression exclude common comorbid anxiety symptoms, and those for anxiety disorders exclude depressive symptoms. However, diagnostic criteria are not the same as clinical presentations. Some somatic symptoms that can occur with both depression and anxiety are outlined in Box 2.

It may be necessary to see a patient on several occasions to delineate his or her problems. If features of anxiety or depression are identified, features of the other disorder should always be sought. For example, if a patient is depressed, a clinician should ask, “With this illness, have you had symptoms like restlessness, irritability, impulsivity, palpitations [or other anxiety symptoms]?”. If a patient is anxious, ask, “With this illness, have you had symptoms like feeling sad or numb, slowed up, loss of energy, a sense of hopelessness [or other depressive symptoms]?”.

As treatments differ if the patient has bipolar depression, ask, “Have you ever had a period of time when you felt ‘up’ or ‘high’ or so full of energy, or irritable or angry, or full of yourself that you got into trouble, or that others thought you were not your usual self?” Bipolar disorder commonly presents as depression, even though the illness is defined by intermittent periods of mood elevation. Anxiety is common in patients with bipolar depression. Manic features that can mimic anxiety include decreased and restless sleep, distractibility, racing thoughts, irritability and agitation.

Despite data suggesting that the accuracy of depression recognition by non-psychiatrists is low,24 the best remedy for this is not clear. Rating scales for depression and anxiety can be helpful, although most are designed to assess the severity of an already diagnosed illness rather than to make a new diagnosis.

A screening instrument that can be used to make a diagnosis and distinguish between illnesses, including unipolar and bipolar depression, is the Mini International Neuropsychiatric Interview.25 Other instruments that can be used to identify the severity (but not the type) of depression include the Kessler Psychological Distress Scale,26 the Hospital Anxiety and Depression Scale,27 or the 12-item Somatic and Psychological Health Report questionnaire.28 The Clinical Global Impression scale is a simple seven-point scale that can be used to monitor progress of treatment response.29

In the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial, about half of the patients with major depressive disorder also had clinically meaningful levels of anxiety.30 Remission was significantly less likely and took longer in the 53% of patients with anxious depression than in those with non-anxious depression. Side-effect frequency, intensity and burden, as well as the number of serious adverse events, were significantly greater and outcomes worse in the group with anxious depression.

There is a paucity of data on the treatment of patients with comorbid depression and anxiety, and clinical practice is therefore determined by treating individual anxiety and depressive disorders. There are also online resources that can help, including the Australian National University’s MoodGYM (http://www.moodgym.anu.edu.au) and E-couch (http://www.ecouch.anu.edu.au), Swinburne University’s Anxiety Online (http://www.anxietyonline.org.au), and information sites from beyondblue (http://www. beyondblue.org.au), Youthbeyondblue (http://www. youthbeyondblue.com) and the Black Dog Institute (http://www.blackdoginstitute.org.au).

Where the anxiety and depression stem from a general medical condition or complications of pharmacotherapy, clinicians should treat the medical disorder and review and adjust pharmacotherapy as necessary. For most patients, there is no obvious medical aetiology.

Treatments for anxiety and depression can have substantial elements in common (Box 3). Initial steps are making the diagnosis, explaining symptomatology, and providing hope. Psychosocial interventions, including clinical support, education and rehabilitation, are valuable. For patients with mild to moderately severe depression and anxiety, structured psychological treatments, available under the federal government’s “Better Access” scheme, will often suffice.31

For patients with more severe illness or those who do not respond to psychological interventions, pharmacotherapy is indicated. Pharmacotherapy particularly decreases over-activity of limbic structures of the brain (bottom-up effect), whereas psychotherapy tends to increase activity and recruitment of frontal areas (top-down effect).32

As most effective treatments for depression also have useful anti-anxiety effects, a pragmatic approach is to begin by treating the depression. Residual specific anxiety disorder symptoms can then be treated, with most responding to psychological interventions rather than additional pharmacotherapy.

Comorbid depression and anxiety are common and affect up to a quarter of patients attending general practice. Screening for comorbidity is important, as such patients are at greater risk of substance misuse, have a worse response to treatment, are more likely to remain disabled, endure a greater burden of disease, and are more likely to use health services in general. There are effective treatments for specific disorders, but a paucity of data about treatment for anxiety and depression comorbidity. More than a third of patients with a mental disorder do not seek treatment, and almost half are offered treatments that may not be beneficial. This suggests the need for further public awareness and professional education that can enhance clinical practice, promoting better mental health outcomes.

3 Flow chart for treating depression and anxiety

Depression and dementia are common syndromes in older people and are usually managed by general practitioners. Comorbidity compounds the impact on patients, carers and health services.1–3 Yet, the relationship between the two is complex — features overlap and each seems to be a possible risk factor, symptom or consequence of the other. Thus, identification and effective management of depression in people with dementia remains a challenging task in clinical practice.

This article provides a clinically oriented selective review of current knowledge about depression in dementia, specifically aimed at primary care practitioners. Practice pearls are provided (Box 1). Although the literature overwhelmingly focuses on Alzheimer disease, here we also discuss mild cognitive impairment and other types of dementia. Mild cognitive impairment refers to a clinical status where a patient performs below norms on cognitive tests but does not have dementia. Either the patient or someone who knows him or her well should have noticed a change from premorbid cognitive function. Statistically, people with mild cognitive impairment are at increased risk of developing dementia over time, although the individual risk can vary significantly.4

Reported rates of depression in dementia vary substantially, depending on the population sampled, means of assessment and definition of caseness. Overall, most well conducted population-based studies report prevalences between 8% and 30%.1,5,6 In hospitalised patients and nursing home residents, the prevalence may be over 40%.1,5 Variance in prevalence estimates is greater in studies of mild cognitive impairment.6 A recent review of depression in mild cognitive impairment found median proportions of 44% in samples of hospital-based patients and 16% in community-based samples.1 The limited studies investigating depression in people with vascular dementia, Lewy body dementia or dementia associated with Parkinson disease suggest that depression may be more common in these syndromes than in Alzheimer disease.1

Reported risk factors include female sex, earlier age at dementia onset,7 a past history of depression or emotional problems,7,8 and recent losses.8 There is some evidence that, contrary to conventional wisdom, stage of dementia and insight into the diagnosis have little impact on depression rates.1,9

The consequences of depression in people with dementia include increased burden on patients and caregivers,10 increased caregiver depression,10 exacerbation of cognitive and functional decline,3,11,12 more dementia-related behavioural disturbance,13 poor outcomes from other medical or surgical interventions,3,14 earlier admission to nursing homes, and increased mortality.12,15–17

There is substantial academic interest in the relationship between depression and dementia, particularly regarding the direction of causality and possible disease mechanisms. There is now reasonable evidence that a history of depression is a risk factor for developing dementia, particularly when the depression occurs early in life or is severe.18 Conversely, incident depression occurring temporally close to the onset of cognitive impairment may represent very early or prodromal symptoms of dementia itself.1,19–21

A neuropathological mechanism common to both conditions, such as cerebrovascular disease or hippocampal atrophy, has been suggested.22–26 However, as with most psychiatric illness, the literature does not consistently support any “neat” pathophysiological hypothesis. It is likely that a complex intertwining of multiple factors is variably involved in different individuals.

Depression in people with dementia remains underdiagnosed, particularly in residential care settings.27,28 Disease, patient, clinician, family and system-level factors may all hinder accurate assessment.

Depression may present differently in people with comorbid dementia, particularly when the dementia is advanced. Although typical major depressive disorder (MDD) is seen, the clinical picture often lacks prominent sadness, hopelessness and guilt and is, overall, less severe.29 Anxiety, psychomotor retardation, loss of energy or appetite, anhedonia, irritability, delusions and hallucinations may all be more common (and prominent) than in people without dementia. These symptoms can also be part of the dementia itself or suggest delirium.3,24,30 Unexpected or rapid change in mood, cognitive deterioration, or increased behavioural and psychological symptoms of dementia (eg, disinhibition, agitation, anxiety or aggression) may be the only indicators of superimposed depression.13

Conventional techniques and instruments for assessing depression may not be reliable or practicable in people with dementia. When obtaining a patient history, patients may underreport, and their carers overreport, symptoms of depression.10 As dementia advances, cognitive and communication difficulties may also hamper assessment of the intrapsychic symptoms of depression, including subjective mood states, thoughts about oneself or others, and general outlook for the future. The use of standardised questionnaires is helpful only if they are validated in this population, which is not the case for many of the common depression rating scales.31

A useful scale is the Cornell Scale for Depression in Dementia (CSDD; Box 2).32 There are also modified Diagnostic and statistical manual of mental disorders, 4th edition (DSM-IV) criteria for MDD that include the diagnosis of Alzheimer disease.33 In Australia, the CSDD is incorporated into the Aged Care Funding Instrument, which is used to determine the funding allocated to aged care facilities for individual residents. However, due to limited time and staff in such facilities, attention to completing the CSDD is often cursory, and the results rarely alter management for individual residents. Moreover, serial evaluations are uncommon.28 Regularly combining use of rating scales with a thorough history would improve detection rates.

Obtaining as much history as possible regarding mood and behavioural symptoms from the patient and carers, such as family or residential care staff, is essential, particularly when communication difficulties are evident. This should be repeated at regular intervals (eg, 6-monthly) or if there is any evidence of a significant change in behaviour.28,34,35 It is important to enquire about a past history of depression and past treatments for depression, as well as family history, medical history, and drug and alcohol history. Physical illnesses, prescription medications, functional disability, social isolation, life stressors, bereavement and other losses are also important, as they may increase vulnerability to depression.1,29

Where there is clinical suspicion of depression (eg, change in mood or any difficult-to-explain change in behaviour), definitive treatment is usually worthwhile, especially when communication difficulties due to dementia act as a barrier to obtaining a detailed history.

Management of a patient with depression and dementia should encompass biological, psychological, social, cultural and spiritual factors, particularly those that may have precipitated or may be perpetuating the depression.

A focused physical examination and investigations are essential to exclude any treatable medical cause for low mood. Such causes are more likely to be present in older people and may include physical illness and prescription medications. Common medical causes of depressed mood are listed in Box 3, and important screening investigations for an organic cause are listed in Box 4. Patients also frequently present with physical illness and depressed mood where there is a known association, but it is neither causal nor reversible (eg, depression and ischaemic heart disease). Overall physical health should be reviewed and optimised, and the depression should be specifically treated. Addressing vascular risk factors is particularly important.

Risks associated with depression in people with dementia must be considered throughout, as they may determine the type, location and urgency of treatment. Every patient with depression should be asked directly but sensitively about suicidal thoughts, plans and intent (Box 5). Suicide risk also depends on available protective supports, supervision and the patient’s access to means. Other risks include harm to others or to relationships due to agitation, aggression or irritability; loss of accommodation or a move to institutional care; and medical compromise. Generally, most depression in people with dementia can be managed in the community.

Depression is common in people with dementia, and the relationship between mood and cognitive symptoms is complex. The nature of this relationship, disease stage, and environmental and clinician factors all contribute to underdiagnosis and undertreatment. Although evidence regarding antidepressant treatment is limited and equivocal, there is no cause for therapeutic nihilism. Organic causes of depressed mood should be excluded, physical health optimised, and medications, where possible, rationalised. Best management includes individually tailored psychosocial strategies and, in moderate to severe depression, judicious use of antidepressant medication with a “start low, go slow” approach. Regular review and effective engagement of carers are both essential and may have significant positive impacts for patients and carers. Difficult-to-treat or complex depression is an indication for specialist referral.