Ignoring atypical symptoms of depression in common scales

Abstract

BACKGROUND

Atypical depression is an important indicator of a high risk of bipolar disorder and a genetic predisposition to immunometabolic traits.

AIM

To analyze common depression assessment scales for their inclusion of items related to atypical symptoms such as mood reactivity, hypersomnia, increased appetite (or weight gain), leaden paralysis, and interpersonal sensitivity.

METHODS

A search for English-language articles was conducted without time restrictions in the MEDLINE and Russian Science Citation Index databases using the following keywords: “depression” OR “bipolar depression” AND “scales” OR “questionnaires”. The analytical method used in this review involved a descriptive analysis of the included studies.

RESULTS

After reviewing studies on the validation of depression assessment scales, we found that only a small number include items addressing both increases and decreases in appetite or weight, as well as variations in sleep duration. Moreover, only a few studies have evaluated mood reactivity, leaden paralysis, and interpersonal sensitivity. The most well-developed scale that considers all aspects of atypical and non-atypical depressions is the Inventory of Depressive Symptomatology.

CONCLUSION

Ignoring atypical symptoms in common scales can lead to underestimation of depression severity and inaccuracies in evaluating therapy effectiveness in clinical trials, as well as hinder fundamental research aimed at finding biomarkers.

Key Words: Atypical depression; Bipolar disorder; Mood reactivity; Depression assessment scales; Inventory of Depressive Symptomatology

Core Tip: Common depression rating scales do not fully address atypical symptoms (e.g., mood reactivity, hypersomnia, hyperphagia, leaden paralysis, and interpersonal sensitivity), potentially underestimating atypical depression severity. Only three scales cover all five atypical features, with the Inventory of Depressive Symptomatology offering the most comprehensive assessment. Incorporating atypical symptoms is crucial for accurate severity evaluation and advancing both clinical and basic research on depression.

INTRODUCTION

One of the key concepts in psychometrics is the latent construct - a hypothetical or theoretical entity (for example, a level of depression or anxiety) that cannot be directly observed or measured[1]. Instead, it is “hidden” (latent) and becomes apparent through a set of indirect signs or indicators (e.g., responses to questionnaire items). The widespread use of latent constructs has contributed to the replication crisis, wherein many research findings prove non-replicable or contradictory[2]. On the one hand, latent constructs have enabled the study of complex mental phenomena; on the other hand, their excessive reliance on mathematical methods has produced “déjà-variables” - cases in which different terms describe the same phenomenon, or a single term is applied to different conditions[2]. In the absence of objective methods for diagnosing mental disorders, the “gold standard” in psychiatry is verifying the diagnosis using the operationalized criteria described in the Diagnostic and Statistical Manual of Mental Disorders (DSM) and the International Classification of Diseases[3,4]. Thus, through consensus in psychiatric classifications, we have established what is considered a mental disorder, and consequently, what can be regarded as a conventional latent construct.

Depression is one of the most common mental disorders in psychometric research[5]. At the same time, its clinical heterogeneity may reflect different biological and causal mechanisms, with growing evidence indicating that various subtypes of depression are distinctly associated with biological risk factors, disease course, and treatment response[6]. Nevertheless, in developing psychometric scales, DSM and International Classification of Diseases specifiers for depression - which point to its different subtypes - are seldom incorporated. This omission represents a significant limitation of such scales, given that their items are initially unable to fully capture the latent construct of depression.

One of the most overlooked subtypes of depression in psychometric scales is atypical depression[7]. For example, the Montgomery-Asberg Depression Rating Scale and the Hamilton Depression Rating Scale, which are used in over 90% of randomized clinical trials of antidepressants for unipolar depression[8], only take into account decreased appetite/weight and sleep[9,10]. Thus, if a depressed individual overeats every day, gains 4 kg in a month, and sleeps 15 hours a day, they would receive 0 point on the appetite/weight and sleep items when using these scales to assess overall depression severity. Consequently, the severity of atypical depression is artificially underestimated compared to patients with the melancholic subtype[11,12].

At the same time, atypical depression is the most common form of depression among outpatients[13]. This subtype is associated with an earlier onset, female sex, and more severe and recurrent depressive episodes[14], and it demonstrates a clear genetic predisposition toward immunometabolic traits and alcohol use[15]. Moreover, atypical depression occurs significantly more often in bipolar disorder - hypersomnia and increased appetite may be present in up to 70% of patients - and can serve as an important distinguishing feature from unipolar depression[16,17]. Nevertheless, in randomized clinical trials of depression in bipolar disorder, scales such as the Montgomery-Asberg Depression Rating Scale and the Hamilton Depression Rating Scale are used just as frequently as they are for unipolar depression[18-22].

The purpose of this review is to analyze the items of depression severity rating scales to determine whether they include or exclude such atypical depression symptoms as mood reactivity, hypersomnia, hyperphagia, leaden paralysis, and interpersonal sensitivity, as well as to identify the most suitable scale for both basic and clinical research on depression.

MATERIALS AND METHODS

Search strategy

A search for English-language articles was conducted without time restrictions in the MEDLINE and Russian Science Citation Index databases using the following keywords: “depression” OR “bipolar depression” AND “scales” OR “questionnaires”. In addition, a manual search was performed by reviewing reference lists of studies on psychometric scales for assessing depression severity.

Eligibility criteria

The inclusion criteria were as follows: (1) Original English-language studies on the development and validation of scales for assessing depression severity; (2) Scales validated in patients with a confirmed diagnosis of major depressive disorder or bipolar disorder type I or II; and (3) Articles and the scales themselves were available in open access. Studies were excluded if: (1) They involved validating an original scale in a specific population; (2) They were conducted among the general population or among patients with mental disorders other than mood disorders; or (3) It was not possible to evaluate the structure of the scale’s items.

Review strategy

From the included articles, the following data were extracted: The title, authors, year of publication, country, method of scale completion, assessment of appetite/weight and sleep changes, mood reactivity, leaden paralysis, and interpersonal sensitivity. Data collection was performed independently by two authors (Yakovleva YV, Shchepkin ES) and confirmed by the other two reviewers (Kasyanov ED and Mazo GE).

Statistical analysis

The analytical method used in this review involved a descriptive analysis of the included studies. In the original studies on the development and validation of depression severity rating scales, we identified indicators for changes in appetite/weight and sleep, as well as the presence or absence of other atypical depression symptoms (mood reactivity, leaden paralysis, and interpersonal sensitivity). Additionally, we recorded the year and country of publication. The extracted data were compiled into a table, and no statistical methods were employed for data analysis.

RESULTS

A total of 31 psychometric scales for assessing depression severity were included in the review, of which 8 (25.8%) were presented in various updated or shortened versions (Table 1[23-56]). The period of their initial validation spans 1960-2014 (from the Hamilton Depression Rating Scale to the Clinically Useful Depression Outcome Scale - Mixed Features Specifier). The majority of these instruments were developed in the United States (22 scales, 71%), followed by the United Kingdom (6 scales, 19.4%) and Australia (5 scales, 16.1%), with some overlap in country authorship. Only 8 (25.8%) scales require clinician administration, and among these, two (the Inventory of Depressive Symptomatology and the Quick Inventory of Depressive Symptomatology) also have self-report versions.

Table 1 Assessment of Atypical Depression Symptoms in Common Scales.

Number	Scale name	Year	Country	Appetite	Weight	Sleep	Reactive mood	Heavy, leaden feeling	Interpersonal sensitivity	SR/CA	Ref.
1	ADDS	1993	United States	Increased	Increased	Increased	+	+	+	SR	[23]
2	BADS-SF	2011	United States	-	-	-	-	-	-	SR	[24]
3	BDI-1А	1961	United States	Decreased	Decreased	Decreased	-	-	-	SR	[25]
	BDI-2	1996	United States	De/increased	-	De/increased	-	-	-	SR	[26]
4	BDRS	2007	Australia	De/increased	-	De/increased	-	-	-	CA	[27]
5	BPRS	2013	Switzerland	-	-	-	-	-	-	CA	[28]
6	BSDS	2005	United States	-	Increased	Increased	-	-	-	SR	[29]
7	CDI	1988	Australia	-	-	-	-	-	-	SR	[30]
8	CDRS	1979	United States	De/increased	-	Decreased	-	-	+	CA	[31]
9	CES-D	1977	United States	Decreased	-	-	-	-	+	SR	[32]
	CES-DC	1986	United States	Decreased	-	Decreased	-	-	+	SR	[33]
10	CUDOS	2008	United States	De/increased	-	De/increased	-	+	-	SR	[34]
	CUDOS-M	2014	United States	De/increased	-	De/increased	-	+	-	SR	[35]
11	DASS 21	1995	Australia	-	-	-	-	-	-	SR	[36]
	DASS 42	1995	Australia	-	-	-	-	-	-	SR	[36]
12	DID	2004	United States	De/increased	De/increased	De/increased	-	-	-	SR	[37]
13	DSRSC	1987	Australia, United Kingdom	-	-	-	-	-	-	SR	[38]
14	EPDS	1987	United Kingdom	-	-	Decreased	-	-	-	SR	[39]
15	GDS	1982	United States	-	-	-	-	-	-	SR	[40]
16	HADS	1983	United Kingdom	-	-	-	-	-	-	SR	[41]
17	HDRS - 17/21	1960	United Kingdom	-	Decreased	Decreased	-	-	-	CA	[10]
18	HANDS	2000	United States	Decreased	-	Decreased	-	-	-	SR	[42]
19	IDS-C/SR	1986	United States	De/increased	De/increased	De/increased	+	+	+	CA and SR	[43]
20	KADS	2003	Canada	-	-	Decreased	-	-	-	SR	[44]
21	LIDAS	2017	Netherlands	De/increased	De/increased	Decreased	-	-	-	SR	[45]
22	MADRS	1979	United Kingdom, Sweden	Decreased	-	Decreased	-	-	-	CA	[9]
23	MDI	2001	Denmark	De/increased	-	De/increased	-	-	-	SR	[46]
24	MFQ-Questionnaire - Adult Self-Report - Long	1993	United States, United Kingdom	De/increased	-	De/increased	-	-	+	SR	[47-49]
	MFQ - Adult Self-Report - Short			-	-	-	-	-	+	SR
	MFQ - Child Self-Report - Long			De/increased	-	De/increased	-	-	+	SR
	MFQ - Child Self-Report - Short			-	-	-	-	-	+	SR
25	PHQ-9	2001	United States	De/increased	-	De/increased	-	-	-	SR	[50]
26	QIDS-C/SR	2004	United States	De/increased	De/increased	De/increased	-	-	-	CA and SR	[51]
27	SCL-90	1973	United States	De/increased	-	Decreased	-	+	+	SR	[52]
28	SDQ	2014	United States	De/increased	De/increased	De/increased	+	+	+	SR	[53]
29	SDS - Zung	1965	United States	Decreased	Decreased	Decreased	-	-	-	SR	[54]
30	SIGH-SAD	1988	United States	De/increased	De/increased	De/increased	-	+	-	CA	[55]
31	WSAS	1988	United States	De/increased	De/increased	De/increased	-	-	+	SR	[56]

SR: Self-report; CA: Clinician administration; ADDS: Atypical Depression Diagnostic Scale; BADS-SF: Behavioral Activation for Depression Scale - Short Form; BDI: Beck Depression Inventory; BDRS: Bipolar Depression Rating Scale; BPRS: Brief Psychiatric Rating Scale; BSDS: Bipolar Spectrum Diagnostic Scale; CDI: Children’s Depression Inventory CDRS - Children’s Depression Rating Scale; CES-D: Centre for Epidemiological Studies - Depression Scale; CES-DC: Center for Epidemiological Studies Depression Scale for children; CUDOS: The Clinically Useful Depression Outcome Scale; CUDOS-M: Clinically Useful Depression Outcome Scale - mixed features specifier; DASS: Depression and Anxiety Stress Scales; DASS: Depression and Anxiety Stress Scales; DID: Diagnostic Inventory for Depression; DSRSC: Depression Self-Rating Scale for children; EPDS: Edinburgh Postnatal Depression Scale; GDS: Geriatric Depression Scale; HADS: Hospital Anxiety and Depression Scale; HDRS: Hamilton Depression Rating Scale; HANDS: Harvard National Depression Screening Scale; IDS-C/SR: Inventory of Depressive Symptomatology; KADS: Kutcher Adolescent Depression Scale ; LIDAS: Lifetime Depression Assessment Self-report; MADRS: Montgomery-Asberg Depression Rating Scale; MDI: Major Depression Inventory; MFQ: Mood and Feelings Questionnaire; PHQ-9: Patient Health Questionnaire - 9; QIDS-C/SR: Quick Inventory of Depressive Symptomatology; SCL-90: Symptom Checklist - 90; SDQ: Symptoms of Depression Questionnaire; SDS: Self-Rating Depression Scale; SIGH-SAD: Structured Interview Guide for the Hamilton Depression Rating Scale - Seasonal Affective Disorder version; WSAS: Weinberg Screen Affective Scale.

According to our analysis, 15 (48.4%) scales assess both decreased and increased appetite; 4 (12.9%) only evaluate decreased appetite, while 1 (3.2%) focuses solely on increased appetite. In another 11 (35.5%) scales, appetite changes are not assessed at all. As expected, body weight changes appear even less frequently: 7 (22.6%) scales measure both weight loss and gain, 3 (9.7%) consider only weight loss (including the first version of the Beck Depression Inventory from 1961, whereas this parameter was excluded in its 1996 revision), and 2 (6.5%) track only weight gain. The remaining 19 (61.3%) scales do not consider weight changes at all. Regarding sleep, bidirectional changes (increase/decrease) are addressed by 12 (38.7%) scales, only reduced sleep by 9 (29%), and increased sleep alone by 2 (6.5%). In 8 (25.8%) scales, sleep disturbances are not taken into account at all.

Analysis of other atypical symptoms showed that mood reactivity is included in only 3 (9.7%) scales, leaden paralysis in 7 (22.6%), and interpersonal sensitivity in 12 (38.7%). Meanwhile, only three instruments (the Atypical Depression Diagnostic Scale, the Symptoms of Depression Questionnaire, and the Inventory of Depressive Symptomatology) account for all five symptoms of atypical depression; the latter two also consider decreased appetite/weight and sleep. Of note, two of the most commonly used scales in clinical research (the Montgomery-Asberg Depression Rating Scale and the Hamilton Depression Rating Scale) focus solely on decreased appetite/weight and sleep, whereas the widely used Patient Health Questionnaire-9 covers increased appetite and sleep but omits other atypical features. Furthermore, 6 (19.4%) scales, all of which are self-report instruments, do not address any changes in appetite/weight or sleep, nor do they include other atypical depression symptoms.

DISCUSSION

In this exploratory analysis of widely used psychometric scales for assessing atypical depression symptoms, we identified several key points. First, only three scales out of the entire set evaluate all five atypical depression symptoms. Second, although mood reactivity is specified as a mandatory criterion in the DSM, most scales place their primary emphasis on changes in appetite and sleep; thus, mood reactivity appears in only three scales, each of which also assesses the other atypical symptoms. Third, weight changes are evaluated less frequently than appetite changes, and weight gain is taken into account in only one-quarter of the scales. Fourth, most scales do not fully consider other atypical symptoms such as leaden paralysis and interpersonal sensitivity, limiting their application in clinical and research settings - particularly when accurate measurement of atypical depression severity is required.

Notably, the Inventory of Depressive Symptomatology stands out as the only scale capable of being used both by clinicians and in self-report format, while simultaneously covering all five atypical depression symptoms[44]. It is worth mentioning that this scale was developed before the introduction of DSM-IV in 1994, which first included a specifier for atypical depression[57]. However, its design was based on the Research Diagnostic Criteria, thus encompassing a wide range of symptoms used to differentiate among atypical, melancholic, and anxious depression[58]. Through factor analysis, four factors were established in the Inventory of Depressive Symptomatology: Mood/cognition (anhedonia, sadness, self-esteem, suicidal ideation); anxiety/hypochondriasis (insomnia, psychomotor changes, somatic complaints, panic/phobias); endogenous symptoms (decreased weight/appetite, insomnia, diurnal mood variation); and atypical symptoms (increased weight/appetite, hypersomnia)[44]. The authors specifically noted that all items have equal weight, given the lack of theoretical or clinical reasons to prioritize certain items in the overall score. Therefore, considering that the Inventory of Depressive Symptomatology offers a more comprehensive description of both atypical and non-atypical depressions and can be utilized by clinicians, it appears to be a more suitable tool for assessing depression severity in both clinical and basic research.

At the same time, the extremely limited attention given to mood reactivity in most analyzed scales is likely linked to prolonged debates about the clinical validity of atypical depression and its diagnostic criteria. For instance, introducing mood reactivity as a mandatory criterion for the DSM-IV atypical depression specifier prompted substantial scholarly debate[13]. Moreover, most primary studies on atypical symptoms were conducted in samples of patients with unipolar depression rather than in those with bipolar disorder, despite these symptoms being significantly more prevalent in the latter group[16,59,60]. Follow-up research on mood reactivity underscored its greater importance specifically in bipolar II disorder: Atypical symptoms were significantly more common in patients exhibiting mood reactivity, whereas in unipolar depression, mood reactivity was not associated with any atypical symptom[60]. In this regard, Benazzi[59] proposed including mood reactivity in the DSM-IV atypical depression specifier only for bipolar II disorder, but not for unipolar depression. Another study of diagnostic criteria for atypical depression found that a non-hierarchical definition of atypical depression regarding mood reactivity provides the same valid characterization of the syndrome as the existing hierarchy based on mood reactivity as a mandatory criterion[61]. Consequently, Angst et al[61] considered retaining all five DSM criteria for atypical depression but without obligating mood reactivity. However, with the release of DSM-5 in 2013, the specifier for atypical depression remained unchanged[3].

In another study, Angst et al[62] demonstrated that truncated definitions of atypical depression - utilizing 2 of 3 (leaden paralysis, increased appetite, hypersomnia) or 2 of 2 (increased appetite, hypersomnia) neurovegetative symptoms - showed the strongest association with female gender, bipolar disorder, and a family history of mania. While such truncated definitions of atypical depression may prove useful in real-world clinical practice for distinguishing unipolar from bipolar depression, it is currently unknown whether they more accurately reflect overall depression severity compared to assessing all five atypical depression criteria. Therefore, despite the lack of comprehensive data, mood reactivity remains one of the core symptoms of atypical depression, and measuring its severity in the context of overall depression severity is a critical subject of clinical research that requires appropriate psychometric scales.

From a methodological standpoint, further progress in the creation and refinement of scales depends on prospective research into how depression and its subtypes evolve over time. This is essential for developing a standardized latent construct of depression, for which new and existing scales can be adapted. Furthermore, another crucial direction involves differentiating the weighting of individual symptoms and creating more nuanced instruments that account for both the mandatory criteria of depression and possibly more specific (or, conversely, less specific) features of its subtypes. Hence, the heterogeneity of depression necessitates including all of its manifestations in the design of psychometric scales. Moreover, to enhance the accuracy of gauging depression severity, a shift away from the paradigm that “all depressive symptoms are equal” is needed. Core symptoms of depression and its subtypes, which mediate illness severity, should be given greater weight than supplementary symptoms. Such a hierarchy of depression symptoms can be developed using Item Response Theory methodologies and machine learning.

This article has several limitations. First, we did not conduct a systematic review that included a methodological quality assessment. Second, our focus was exclusively on the DSM’s atypical depression specifier criteria, rather than the principal diagnostic criteria or those for other depression specifiers. Third, we did not compare how closely each scale’s item wording aligns with one another or with the DSM’s atypical depression criteria.

CONCLUSION

Our findings indicate that most commonly used psychometric scales for assessing depression severity devote insufficient attention to atypical symptoms. Only three instruments (the Atypical Depression Diagnostic Scale, the Symptoms of Depression Questionnaire, and the Inventory of Depressive Symptomatology) capture all five features of atypical depression, including mood reactivity. However, among these, the Inventory of Depressive Symptomatology stands out for its versatility (it is available both as a clinician-administered and a self-report version). Future progress in psychometrics will require prospective research on the course of depression, clarification of the weighting of individual symptoms, and validation across diverse clinical groups of patients with depression. Such endeavors may lead to more accurate and sensitive tools for assessing overall depression severity and, in particular, atypical depression.