Unifying the diagnosis of mood disorders

Introduction

The validity of mood disorders diagnoses awaits the discovery of a definitive ‘biomarker’. Current diagnoses rely upon the clustering of symptoms that characterise the many presentations of mood disorders. None of these symptoms are pathognomonic, that is, uniquely characteristic of mood disorders, and most are merely features, which in and of themselves have little specificity, such as changes in appetite and patterns of sleep. However, some signs and symptoms such as anhedonia and guilt probably carry more weight than others in signifying depression, and similarly in mania, flight of ideas and prolixity, or markedly increased energy, are perhaps more specific. Nevertheless, a clinical conundrum remains; how do we reliably distinguish clinically significant mood disorders from the normal vicissitudes of day-to-day emotions, and how can we accurately diagnose them so that they meaningfully inform treatment.

A key initial step has been to carve out distinct diagnostic categories from the real-world clinical presentations that usually manifest as symptom dimensions. Prototypical cases, characterised by dramatic changes in mood, for instance the abrupt onset of manic symptoms, or a rapid descent into melancholia, are easy to detect and recognise as pathological, and clearly warrant treatment. However, in the majority of cases, the emergence of mood disorders has an insidious onset in which symptoms emerge over an extended period of time – often fluctuating of their own accord or in response to compensatory and adaptive processes. Such flux makes it difficult to determine exactly when clinical symptoms reach a level that signifies ‘pathology’ and the development of a case.

To address this problem, the first and most obvious step is to identify a threshold and, over the years, a variety of approaches have been employed. For example, in clinical trials a number of scales that rate individual symptoms and sum the scores (e.g. MADRS and YMRS) are used to define mood disorders on the basis of overall severity (Montgomery and Åsberg, 1979; Young et al., 1978). Others place emphasis on particular kinds of clinical features (e.g. symptoms of melancholia) or attempt to gauge the degree of impairment an illness confers (e.g. the functional limitations associated with a mood disorder). However, none of these measures are ideal, and yet, many of these approaches form the basis of our current taxonomy that yields ill-defined mood disorder diagnoses. This is partly because mood rating scales were developed primarily to select patients for clinical trials – and in this context they were not being used as diagnostic instruments per se. Instead they were used to ensure homogeneity of the sample chosen for study and for this purpose the overall severity score of symptoms was sufficient. Furthermore, it is important to note that in clinical practice the inter-rater reliability of many of these rating scales was relatively poor; yet they remain a core part of patient selection – their inclusion being enabled to some extent by drug approval bodies wanting increased validity.

Alongside inquiry into the phenomenology of mood disorders, attempts at developing a psychopathology that increases understanding into its pathophysiology have advanced considerably. Early indicators from studies conducted on the stress axis, genetics and epigenetics, neuroimaging and neurocognitive studies have been promising (Malhi and Mann, 2018). But as yet, no marker (neurological or biological), or groups of such markers have been identified and consequently no objective measures or tests are available in clinical practice. Therefore, in the meantime, there is little choice but to focus on the phenomenology of mood disorders and formulate phenotypes that provide ontological understanding.

In practice, clinicians adopt a pragmatic approach for the evaluation and diagnosis of mood disorders and often adhere to one schema (severity, subtyping) with experienced clinicians often using pattern recognition. These practices are influenced in part by the major classificatory systems because categorical subtypes (e.g. major depression, persistent depression and different forms of bipolar disorder) are emphasised in Diagnostic and Statistical Manual of Mental Disorders (DSM; capitalising on comorbidity) and severity (e.g. mild, moderate and severe depression) features prominently in International Classification of Diseases (ICD). In addition, diagnoses are usually made cross-sectionally, that is, over a relatively short period of time, even though a longitudinal perspective is widely advocated given the intrinsically recurrent nature of mood disorders.

While neither of these approaches (severity and subtyping) is ideal, using them together may increase diagnostic specificity. To sharpen this further, we suggest adding another perspective, namely, the ACE model of classification (Activity, Cognition and Emotion), which has been articulated recently to better capture mixed mood states (Malhi et al., 2018). The ACE model emphasises a dimensional perspective with respect to phenomenology and gives equal importance to activity and cognition alongside emotion – the mainstay of mood disorders.

Diagnostic approaches

When appraising mood disorders, they can be diagnosed and classified according to severity, subtype and the domains in which the symptoms feature.

Severity

Usually, the description of a mood disorder, such as depression, as mild, moderate or severe is a function of the number of symptoms present, as well as a number of other factors, such as degree of distress (that may be more a function of personality style than a true reflection of severity) and the overall functional impairment the individual is experiencing. This may seem crude, but in practice it is a reasonably reliable metric that is straightforward to apply and has the advantage of directly informing management. For example, mild and moderate depressions can be treated within the community using psychological interventions either alone, or in combination with pharmacotherapy. Severe presentations often require more complicated treatment strategies, the engagement of a specialist and sometimes may warrant hospitalisation. This severity-based approach is uncomplicated, but there are some instances in which it can run into difficulties. For example, symptoms such as suicidal ideation and psychotic symptoms do not always align with severity. For instance, by and large, psychotic symptoms are more common in severe depressions and severe mania, but not always, and similarly, suicidal thoughts are often regarded as a proxy for severity, even though, in reality, they often occur in mild and moderate depressions.

In addition, the cut-offs for the different levels of severity (mild, moderate and severe) are difficult to define and depend on the scales that are used, and whether they are self-report, or observer rated. A further complication is that the severity of symptoms can change considerably over the course of a day, as seen with the diurnal variation of mood found in melancholia. In melancholia, the time of day the patient is assessed can markedly affect the rating of severity. Thus, while severity is seemingly straightforward to assess and communicate, it lacks specificity and consistency.

Functionality

Functional change is a key feature of mood disorders, which confer considerable impairment across all aspects of an individual’s life. Functional compromise can be measured globally and also more granularly with respect to particular roles or aspects of mental functioning – and it is an important means of thresholding mood disorders. A functional assessment can also be used alongside other measures to make a diagnosis or gauge its impact or can be used as an adjunct to inform management, once a syndrome has been defined. However, it is important to note that functionality is a complex construct and that in clinical practice it is best assessed longitudinally incorporating information from the individual and as many additional sources as possible (e.g. relatives).

Subtypes

Fundamentally, both depressive and manic syndromes are an aggregation of symptoms that are often accompanied by distress and/or disruptive behaviour. When the functional impact of these changes becomes significant, the individual typically seeks help, and this then leads to the detection, diagnosis and management of their illness. Subtypes of mood disorders, in particular depression, that are characterised by sets of distinctive symptom patterns are useful in this regard as they may facilitate detection. In the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5), the different symptom patterns of subtypes are described as specifiers of the mood episode such as melancholic, psychotic, anxious and atypical (American Psychiatric Association, 2013). In other words, these are subtypes of the mood disorder and they convey a richer clinical picture than a simple overall measure of distress or severity. However, while these subtypes may loosely inform treatment choice, they are not particularly informative regarding outcome. Nevertheless, clinically, it is useful to know which symptoms are most troubling, for example in anxious depression, knowing that anxiety is a prominent feature that helps focus treatment selection. Similarly, the subtyping of depression is useful when psychotic symptoms are present and in these cases treatment with antipsychotic medication is indicated. However, problems remain, as the definitions of subtypes vary (e.g. atypical depression and melancholia are defined differently by different groups around the world) and in practice, the vast majority of mood disorder presentations fail to neatly fit into a particular subtype – limiting the clinical utility of this approach.

ACE model

The ACE model (Malhi et al., 2018) is an alternative approach that groups symptoms into three domains: Activity, Cognition and Emotion. These domains, and the symptoms they contain, are conceptualised as dimensional constructs, rather than categories. Framing symptoms within the ACE model automatically focusses greater attention on the domains of activity and cognition ensuring that these are given equal consideration as mood (which is subsumed within emotion). The overlapping and dimensional aspects of the model also provides a perspective that encourages symptoms, syndromes and ultimately diagnoses to be viewed as a continuum – with disorders themselves forming categories from within a spectrum of phenomenology.

A key advance of the ACE model is that it promotes a more sophisticated paradigm than the simplistic unipolar/bipolar dichotomy which unduly emphasises the polarity of mood. The symptoms and signs in the ACE model exist along a number of intersecting and overlapping dimensions, and this allows for the co-occurrence of symptoms to form admixtures that reflect the reality of clinical practice, as for example, in mixed mood states, which extend beyond the linear unipolar-bipolar axis. By emphasising cognition and activity, the ACE model untethers diagnosis from mood and also allows symptoms from seemingly ‘opposite’ poles of mood disorders to co-occur. This is necessary in order to better capture mixed mood states in clinical practice in which, for example, irritability (a diagnostic symptom of mania) often features strongly alongside anhedonia and low mood (typical symptoms of depression). The ACE model also focusses attention on different symptomatic domains, which are likely to warrant different treatments and different kinds of interventions. For example, changes in the cognitive domain may respond better to psychological interventions – prompting their use earlier and in a more targeted manner.

However, the ACE model does not obviate the difficulty of having to define a threshold that denotes disorder and hence why an amalgamation of a number of approaches is necessary.

A unified approach to diagnosis

Figure 1 illustrates the various dimensions and domains that have been outlined above. It shows that early symptoms likely possess elements of the three ACE domains, and that as symptoms assume form they may gravitate to one or more dimension. Symptoms may also interact and coalesce with other symptoms to produce more complex phenomena, which ultimately form the substrates of clinical syndromes that are then diagnosed as disorders. (Please see detailed explanation in figure legend.)

Figure 1.

The unification of diagnostic approaches for mood disorders.

An integrative approach within the domains of the ACE model allows for the intersection of subtypes, severity and functionality in a manner that provides greater specificity than reliance on any one schema. And by applying these approaches in combination a more meaningful diagnosis can be achieved – that better informs management.

Conclusion

The purpose of achieving a more accurate diagnosis using such a multidimensional approach is to characterise the mood disorder with greater specificity so that illness course and response to treatment can be predicted.

Clearly, any such diagnosis still needs to be nested within a broader formulation that also takes into consideration the unique circumstances of the individual. But by adopting a unified approach we are ensuring that we are making the best use of the information and knowledge that we have at present and that precision management – the current zeitgeist, is at least based on precision diagnosis.