The Bereavement Exclusion for the Diagnosis of Major Depression

Abstract

This paper reviews studies bearing on the validity of the bereavement exclusion for the diagnosis of major depression. It concludes that the exclusion is not supported by the best available data, and the authors propose revisions for Diagnostic and Statistical Manual for Mental Disorders, Fifth Edition.

INTRODUCTION

Mr. A visits your office complaining of five weeks of a persistently low mood, inability to enjoy his usually enjoyable activities, difficulty falling and staying asleep, daytime fatigue, inability to concentrate or make decisions, and poor appetite. He is otherwise in excellent health. Six weeks ago, his wife passed away.

Mr. B visits you with the same complaints and duration as Mr. A. However, his wife is alive and well. Unfortunately, six weeks ago he lost his job, discovered that his life savings had suddenly been lost, and could not pay his mortgage.

Mr. C visits you with the same symptoms and duration as Mrs. A and B. He has lived a charmed life and cannot understand why he feels so bad when everything else is going so well.

Which of these three patients has major depressive disorder (MDD)? The answer, unfortunately, is, “It depends on whom you ask.” If you ask the Diagnostic and Statistical Manual for Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR), you will find that only Mr. B and Mr. C have MDD; Mr. A is diagnosed with the V Code (that is, not psychiatrically ill) bereavement. On the other hand, if you asked the other “official” diagnostic bible, The International Classification of Diseases, Tenth Revision (ICD-10), you will be told that all three men have MDD. Even the “experts” will give you different answers, not always coinciding with the DSM-IV-TR or ICD. For example, Wakefield et al¹ would diagnose only Mr. C with MDD. Frances^2,3 would agree with the DSM-IV-TR and give Mr. B and Mr. C the diagnosis of MDD. Kendler et al,⁴ Zisook et al,⁵ Karam,⁶ Pies,^7,8 Corruble et al,⁹ and Kessing et al¹⁰ likely would side with the ICD-10 and diagnose all three men with MDD.¹¹ Most experts would agree that this important issue deserves further investigation.¹² This paper will briefly discuss the background of the debate, some of the key studies bearing on the two sides, and finally, the authors’ recommendations.

Background

The broader debate. Psychiatric nosology has been accused of being overly inclusive, diagnosing individuals with mental illnesses without due appreciation for cultural diversity and the wide variety of normal human experiences.¹³ At the same time, depriving individuals who have treatable mental illnesses of prompt and appropriate treatment may have profound and life-long consequences. This conundrum may be more evident in individuals with MDD than in many other conditions. For example, the fallacy of misplaced empathy refers to a well-intentioned clinician’s paradoxically missing the diagnosis of MDD, because he or she can “understand” that “anybody” undergoing a serious life stressor—whether becoming disabled, impoverished, terminally ill, humiliated, or bereaved—might be distraught and upset.¹⁴ But it simply does not follow logically that, just because one’s reaction to an event is “understandable,” it cannot be pathological and in many cases severely debilitating. Such misplaced empathy may interfere with accurate diagnosis and much needed treatment. The dilemma lies in knowing where to draw the line along the health-illness continuum.

It is also important to note that individuals who are depressed frequently search for reasons why, and may themselves mistakenly identify events occurring in proximity to the onset of the mood change as causes of the depression.^15,16 Furthermore, being in a depressed state may itself precipitate “dependent” or adverse life events, such as dropping out of school, leaving a job or a relationship, or exposing oneself to risky situations. In effect, depression often “recruits” misadventure, making judgments regarding cause and effect hazardous.⁸ Finally, attributing a depressive episode to a specific “cause,” such as recent bereavement, runs the risk of missing numerous contributing or subsidiary causes, including but not limited to underlying medical or neurological disorders.^8,15,16

The narrow debate. No one wants to pathologize grief, which is a normal, adaptive response to loss. Nor do any enlightened clinicians wish to ignore MDD, when present, or to deny anyone potentially life-changing treatments because the patient’s MDD is misconstrued as “only” grief, or a “normal reaction” to some other severe life event. But where normal grief ends—with its attendant sadness, remorse, and distress—and minor or major depression begins has not been scientifically scrutinized.

The framers of the DSM-III chose to use a mainly “atheoretical” approach to diagnoses, focusing on intensity and duration of symptom patterns and on significant distress or dysfunction, rather than on a presumed etiology for psychiatric disorders. According to DSM-III and its more recent variants, many human experiences are associated with periods of sadness and other stress-related symptoms. These “normal” human experiences should not be diagnosed as major depressive episodes unless criteria are met for severity, duration, and clinically significant distress or impairment.

The single exception to this rule is when the experience is the death of a loved one. Currently, the DSM-IV-TR states that when the symptoms begin within two months of the death of a loved one and do not persist beyond these two months, the diagnosis of major depression should not be made, unless the symptoms (a) are associated with marked functional impairment or (b) include certain “conditional” features, namely morbid preoccupation with worthlessness, suicidal ideation, psychotic symptoms, or psychomotor retardation. The DSM-IV-TR’s rationale is that a depressive state is an expectable and culturally sanctioned response to the death of a loved one and, as such, does not represent a mental disorder.¹²

Much of the impetus for the DSM-III’s introduction of the DSM exclusion came from the ground-breaking studies of Clayton et al^17–21 at Washington University, St. Louis, Missouri. Studying more than 200 widows and widowers, Clayton et al documented the high prevalence of major depressive syndromes occurring during bereavement. Thirty-five percent of widows and widowers met criteria for clinical depression one month after their spouse’s death, and one-third of those remained depressed for at least a year. Because these depressive syndromes tended to be relatively mild, usually dissipated over time and without treatment, and “differed” from clinical depression in several ways, Clayton cautioned against over-diagnosing major depression during the first year of bereavement. More recently, Clayton has advocated retaining the V Code, Bereavement, in DSM-V, but opined that “it may be that the bereavement exclusion for the diagnosis of MDD should be deleted.”²² The next sections of this manuscript consider data arguing for or against maintaining the bereavement exclusion.

Key Studies since DSM-III: the Main Arguments

Keep the exclusion and expand it (Not only does Mr. A not have MDD, but neither does Mr. B). In a secondary analysis of the National Co-morbidity Study (NCS), Wakefield et al¹ suggest that the DSM’s “failure to consider stressful contexts results in false-positive diagnoses; that is, [in] the classification of psychiatrically normal persons as mentally disordered.” Wakefield et al believe that some of these individuals “…are not experiencing a mood disorder, but intense normal sadness in response to bereavement.” Wakefield et al categorized subjects as those with MDD episodes triggered by either bereavement or other loss, and then further divided the groups into “uncomplicated” and “complicated cases.” Uncomplicated cases lacked feelings of worthlessness, suicidal ideation, marked functional impairment or psychomotor retardation, or prolonged duration. The study presented two main findings: 1) MDD after bereavement was far more similar to, than different from, depressive syndromes that occurred after other stressful life events; and 2) subjects categorized as “complicated” cases showed more severe pathology than those with “uncomplicated” cases. From these findings, the authors reached two main conclusions: 1) Bereavement–related depression does not deserve special diagnostic status compared with other life-event triggered depression; and 2) All “uncomplicated,” “triggered depressions” should be exempt from the diagnosis of MDD.

In a subsequent exchange of letters^23,24 Kendler and Zisook described the second Wakefield et al finding—i.e., that “complicated” cases showed more severe pathology than those with “uncomplicated” cases—as “largely tautological.”²⁴ Similarly, while agreeing that bereavement-related depression does not deserve special diagnostic status compared with other life-event–triggered depression, Pies and Zisook²⁵ also found the second finding tautological. They pointed out that the Wakefield et al NCS study did not compare depression “with and without cause;” indeed, fewer than five percent of all cases lacked an identifiable environmental trigger. Nor was the study prospective. Even Wakefield et al themselves acknowledged that “…the accuracy of the … respondents’ self reports of triggering events is unknown; respondents may misremember whether there was an event, or whether the timing of an event was before an episode. Further, they may misattribute the cause of an episode to an event when they were coincidental.”¹

Remove the exclusion (Mr. A, Mr. B, and Mr. C. all have the same disorder: MDD). Several recent reviews and studies have argued that the preponderance of available data suggest the bereavement exclusion can no longer be justified.

In a comprehensive literature review of studies bearing on the question of whether bereavement-related depression is different than nonbereavement-related depression, Zisook and Kendler²⁶ reported that the clinical characteristics, consequences, and course of bereavement-related major depression (BRMD) are similar to those of other, nonbereavement-related MDD. Documented adverse consequences of BRMD include the following: impaired psychosocial functioning; comorbidity with a number of anxiety disorders; and symptoms of worthlessness, psychomotor changes, and suicidality. Symptoms of BRMD are often severe and may be as long lasting as other types of major depression. In addition, BRMD has biological characteristics that reflect similarities with other depressions, such as increased adrenocortical activity, impaired immune function, and disrupted sleep architecture. Moreover, response to antidepressant medications is similar to what would be expected of any nonbereavement- related instances of MDD. Another review,⁵ focusing exclusively on studies evaluating validators of BRMD within months of a loved one’s death and therefore more germane to the question of the DSM bereavement exclusion, found similar results. These reviews conclude that excluding recently bereaved individuals from the diagnosis of MDD, when all other symptomatic, duration, and functional impairment criteria for MDD are met, cannot be justified by the best available data.

Since these reviews, four other studies have been published that provide further evidence supporting the removal of the bereavement exclusion. In the first of to these, Kendler et al⁴ evaluated the empirical validity of the bereavement exclusion by comparing individuals with bereavement-related depressions to those with other stressful life event related MDD in the large, community based, Virginia Twin Study of Psychiatric and Substance Use Disorders (VATSPSUD). In addition, the sample was divided into those meeting criteria for “normal” grief (duration >2 months and absence of suicidal ideation and severe work impairment) and other depressive cases that did not meet normal grief criteria. Like the Wakefield et al study described previously, this study found that bereavement-related depressions were remarkably similar to other, nonbereavement, stress-related depressions, including in the likelihood of lifetime MDD occurring in the co-twin. In addition, the depressions occurring in individuals meeting criteria for “normal” grief were not substantially different from other depressions in intensity, course, comorbidity, or other associated features, providing no support for the special treatment given to bereavement-related depression in the DSM. The investigators concluded that bereavement-related depressions are much more similar to other stress-related depressions than they are to “normal sadness” and that the bereavement exclusion for the diagnosis of MDD cannot be defended by the data.

Using different designs and unique populations, three other recent studies have provided confirmatory data. Karam et al⁶ compared bereavement-related to nonbereavement-related MDD in a large, population-based, prospective study in Lebanon. This study found that the prevalence of the DSM-IV “conditional criteria” was substantial in the total sample (N=685) and did not differ between bereaved and nonbereaved groups. The global symptom profile of depressed individuals and their risk for depressive recurrence was similar in bereaved and nonbereaved subjects, and the duration of illness was actually longer in the bereaved group.

Similarly, Kessing et al¹⁰ explored the Danish Psychiatric Central Research Register to compare characteristics associated with first onset MDD following bereavement (N=26) to MDD following other stressful life events (N=163), as well as to MDD occurring in the absence of life adversity (N=112). They reported that patients who had experienced bereavement did not differ from patients with other stressful life events, or from patients without stressful life events, with respect to sociodemographic variables, clinical characteristics of the depression, psychiatric comorbidity, family history, or response to antidepressant treatment. Their conclusion was that first-episode MDD following bereavement is not substantially different from other kinds of first-episode depression. It should be noted that this study did not separately analyze specific DSM “conditional” criteria for use of the bereavement exclusion. Also, the number of patients with bereavement-related depression was rather small (N=26, 8.6% of the total sample) and the study did not focus specifically on the two-month period following loss of a loved one; rather, the study extended this period up to six months. An important strength of this study, however, was its use of standardized interviews to assess main and comorbid diagnoses, stressful life events, and other variables.

Finally, in a large, case-control, cross-sectional study of a national database in France, Corruble et al⁹ compared three cohorts: 1,521 individuals who met symptom criteria for MDD, but were not diagnosed with MDD because of recent bereavement (“bereavement-excluded patients”); 292 recently bereaved individuals who were diagnosed with MDD (“MDD despite bereavement”); and 1,229 individuals diagnosed with MDD who were not recently bereaved. All of the subjects were assessed using the Montgomery-Asberg Depression Rating Scale (MADRS). The study found that bereavement-excluded subjects were more severely depressed than MDE controls without bereavement and similar to MDE controls with bereavement. Furthermore, two of the conditional symptoms meant to suggest MDD rather than “bereavement”—suicidal ideation and feelings of worthlessness—were actually more common in the bereavement-excluded group than in the group diagnosed with MDD despite bereavement. They argued that their results, like the results of the Zisook,⁵ Kendler,⁴ and Karam⁶ studies, called for a removal of the DSM-IV-TR bereavement exclusion. Furthermore, the authors concluded that use of the DSM-IV bereavement exclusion could “…result in patients failing to be correctly diagnosed…and not getting appropriate treatment.” In a response to the Corruble et al⁹ study, Clayton²² opined that the results were related more to a misapplication of the bereavement exclusion criteria than to an inherent conceptual problem with the bereavement exclusion. Clayton concluded that “If the [bereavement exclusion] criteria are confusing and delegate people seeking treatment after bereavement to a V code, it may be that the instructions are poorly written and that criterion E [“symptoms are not better accounted for by Bereavement”] for major depression should be deleted, but the V code should remain.”²²

Implications and Recommendations for DSM-V

To be uncertain is to be uncomfortable, but to be certain is to be ridiculous.

-Old Chinese Proverb

The preponderance of available data bearing on the validity of the bereavement exclusion for the diagnosis of MDD suggests that new conventions are overdue. Recent bereavement does not “immunize” individuals from MDD when they otherwise meet full symptomatic and duration criteria for major depression. There are no convincing data demonstrating that depression in the context of death of a loved one is fundamentally different from other kinds of depression, though the focus of psychotherapy may of course differ in cases of bereavement-related major depression. In contrast, there is converging, albeit indirect, evidence that the “depressions of bereavement” are no less severe or persistent than other, nonbereavement-related depressions. Eliminating the bereavement exclusion is not meant to “medicalize” grief; nor is it meant to suggest that the vast majority of bereaved individuals—who hurt, grieve, feel sad, may have trouble sleeping, eating, and concentrating—require professional help. Indeed, most bereaved individuals, despite their acute anguish and distress, do regain their emotional footing without treatment. On the other hand, eliminating the bereavement exclusion may facilitate accurate diagnosis and appropriate treatment for those bereaved individuals who do need help.

A modest proposal

Based on the data reviewed here, we concur with the preliminary report from the DSM-V Mood Disorders work group, which indicates their intention to eliminate the “bereavement exclusion” for MDD diagnosis.³¹ At the same time, we support efforts to avoid “medicalization” of putatively transient depressive states or “normal grief”—whether in or not in the context of recent bereavement. We take the position that the current two-week duration criterion for MDD diagnosis is not derived from careful, longitudinal studies of depression outcome and is therefore somewhat arbitrary. We further believe that, in most cases, it is medically inadvisable to base critical treatment decisions on so short an interval of depressive signs and symptoms, which, in our experience, often resolve spontaneously over subsequent weeks. Accordingly, we propose consideration—and ideally, field-testing—of revised duration requirements for MDD, which takes into account both severity and past psychiatric history. Specifically, we propose the following:

• The duration requirement for MDD would be retained at the current (DSM-IV) period of two weeks, when the depression is accompanied by either (1) prominent or frequent suicidal ideation, intentions, or plans; or psychosis (e.g., delusions, accusatory or threatening auditory hallucinations); or (2) presence of DSM-IV “melancholia” (e.g., marked anhedonia, no reactivity of mood, morbid feelings of worthlessness, or prominent psychomotor changes). These guidelines would apply whether or not the depression occurs in the temporal context of significant recent loss or bereavement.

• The diagnosis of MDD may be made after only one week of symptoms if the individual has a well-documented history (e.g., clinical records, reports of family members) of one or more severe MDD episodes with presenting features similar to those of the index (current) episode.

• Absent the markers of severity and risk noted above (e.g., suicidal risk, psychosis, melancholic features, or “high-risk” past history), the diagnosis of MDD should not be made until at least four weeks of relatively persistent signs and symptoms of major depression.

These revised duration criteria may attenuate many of the concerns raised about eliminating the bereavement exclusion for the diagnosis of MDD.² In the context of the “new” duration requirements, we also suggest the following modifications of “bereavement” for DSM-V:

• Eliminate the bereavement exclusion for the diagnosis of MDD.

• “Bereavement” may be retained as a “V” code, i.e., not a mental disorder, and would, by definition, be a condition that does not meet full symptomatic criteria for MDD (as defined above); and is not accompanied by frank suicidal intentions or plans or by psychotic features. However, individuals with “bereavement” often show two or more features of MDD for several weeks or longer (e.g., insomnia, poor appetite, moderate weight loss, fleeting thoughts of “not going on”) after the initial loss. Individuals with bereavement often show phenomenology suggestive of “productive” or adaptive grief, and sometimes may be distinguished from those with major depression partly on this basis (Table 1).^14,27–30 If “bereavement” is retained as a V code, we suggest the DSM-V includes these phenomenological features in the text of the document, though we acknowledge that careful, empirical studies are needed to validate these primarily clinical observations.

• Alternatively, some cases of bereavement associated with significant impairment or incapacity may reasonably be considered a type of adjustment disorder.

• Complicated grief (sometimes termed prolonged grief reaction) refers to severe, prolonged psychic pain, pining and yearning, preoccupation with memories and thoughts of the deceased, and inability to get on with life without the deceased at least six months after the death. It is considered a serious condition requiring intervention and is separate from either (uncomplicated) “bereavement,” as described previously, or MDD.²⁷ The syndrome of complicated grief (i.e., prolonged grief reaction) is being considered for inclusion in DSM-V, and we recommend this be given serious consideration.

Table 1.

Phenomenology of “bereavement” compared with major depression^14,27–30

BEREAVEMENT	MAJOR DEPRESSION/MELANCHOLIA
Emotional connection with significant others preserved (“I-thou” state).	Self-focused, depressed person feels outcast, alienated, alone.
A sense that grief is time-limited; life will eventually be better.	Time stands still; depression feels limitless, never-ending; time itself is experienced as slowed or stopped.
Self esteem and personal potency generally well-preserved; guilt, if present, is focused on what was or was not done for the deceased.	Person experiences self loathing, guilt, low self esteem, sense of personal impotence; guilt is focused on “sins,” or being a worthless, unforgiveable person.
Rarely suicidal; if thoughts of dying are present, they are focused on joining the deceased.	Often suicidal; thoughts of dying focused on not being worthy of living.
Grief is mixed with positive feelings, such as pleasant memories of a lost loved one.	Person lacks positive feelings or memories; may feel ambivalent, conflicted over loss.
Grieving person can be consoled (e.g., by friends, literature).	Person often inconsolable; mood often autonomous, impervious to others.
Dysphoria often experienced in “waves;” circumscribed; often triggered by thoughts, memories of deceased person.	Dysphoria described as diffuse, “always there” (pervasive); person rarely focused on specific person other than self.

Conclusion

Although definitive, longitudinal studies in carefully selected cohorts of depressed patients have yet to be done, we believe that the best available data do not support the validity of the current DSM-IV bereavement exclusion. Most available studies support the conclusion that major depressive symptoms occurring weeks or months after the loss of a loved one do not differ in any important clinical respect from major depressive symptoms following any other loss or from depression occurring in the context of no discernable loss. At the same time, many bereaved persons will show some symptoms of major depression for weeks or sometimes months following their loss, which fall short of MDD criteria. This does not necessarily represent mental disorder. Indeed, normal grief or bereavement should not be “medicalized;” however, neither should bereavement-related major depression be “normalized” merely because one can point to a recent loss. We believe the best way to deal with the potential for “over-diagnosis” of major depression is by modifying the duration criteria for MDD rather than by retaining the bereavement exclusion. Eliminating the bereavement exclusion may allow early diagnosis and treatment of individuals with potentially serious major depressive illness.