Depression or Bereavement? Defining the Distinction

Michael B. First, MD; Ronald W. Pies, MD; Sidney Zisook, MD


April 08, 2011

In This Article

The Last Word

Drs. Zisook and Pies: Once again, we very much appreciate Michael First's thoughtful and spirited rejoinder. Once again, we find ourselves largely in agreement with Dr. First on several key issues. Specifically, we would all agree that MDD is a very broad and heterogeneous category that encompasses both severe, prolonged, melancholic/psychotic episodes at 1 end of the spectrum; and, at the other end, relatively mild, brief, nonmelancholic/nonpsychotic episodes, often associated with loss or other life stressors.

We suspect that all of us also agree on that there are both advantages as well as potential drawbacks of developing "tighter" criteria for MDD, involving either a higher threshold for inclusion (eg, requiring 7 of 9, rather than 5 of 9 symptoms); a longer duration of symptoms in at least some instances; or perhaps both. In this regard, we acknowledge, as Dr. First points out, that the DSM-IV requires only that symptoms "cause clinically significant distress or impairment in 1 or more important areas of functioning" -- not, as we had phrased it, that symptoms "significantly interfere with social/vocational function." However, we consider this to be a problem engendered by the low threshold of the current MDD criteria and not a matter directly relevant to the debate over the bereavement exclusion. Indeed, we believe that most of Dr. First's arguments would be properly and more profitably directed at deficiencies in the current DSM criteria for MDD and not at our proposal for eliminating the bereavement exclusion.

With respect to the so-called conditional criteria now proffered by DSM-IV as a way of differentiating bereavement from major depression (eg, MDD is suggested by persistence of depression for longer than 2 months, marked functional impairment, morbid preoccupation with worthlessness, suicidal ideation, etc), we fully agree with Dr. First that these criteria are based on a "rather limited empirical foundation." In fact, clinical data from several studies referenced in our previous responses find that these features have little specificity or predictive value.

So Is There a Difference?

We believe the fundamental problem in Dr. First's position is the inability to demonstrate, on the basis of even a single clinical study, that symptomatic "MDD" closely following bereavement differs fundamentally from any other instances of MDD, whether related to loss or not. This problem is highlighted in Dr. First's analysis of Dr. Clayton's studies, which is open to both logical and empirical objections. Thus, Dr. First asserts that, "The bereavement exclusion was added to the DSM-III criteria for MDE in recognition of the work of Paula Clayton and others did examining the occurrence of depressive symptoms in widows and widowers shortly after the death of their spouses. They found that it was common for them to develop symptoms that resembled those of a depressive episode..." (our italics).

But the term "resembled" asserts as true precisely what is in dispute. The question is, "In what clinical sense did Clayton's depressed bereaved subjects not have an MDE, but merely something that 'resembled' MDD?" It appears that Dr. First wants to argue that since many subjects in Dr. Clayton's studies showed marked improvement in, or resolution of, their depressive symptoms within a few months of bereavement, they did not have "real" MDD in the first place. However, this assumes that (a) we know with a high degree of confidence what the "true course" of MDD is; and that (b) any constellation of mild-moderate MDD symptoms that do not last at least 8 weeks are not bona fide MDD. We regard the first proposition as questionable, and the second as probably incorrect, based on the limited available data. Yet as we read his remarks, it appears that Dr. First assumes that "real" MDD, spontaneous MDD, or nonbereavement-related MDD must last more than 8 weeks. But there are credible data that argue against this supposition.

The most relevant article addressing this issue is Posternak and colleagues' study.[28] These researchers examined a cohort of patients severely depressed enough to have received treatment for an episode of MDD, and who subsequently experienced a new episode of MDD. The study found that nearly 40% of those not receiving treatment for the new episode remitted spontaneously within 8 weeks. Specifically, in the 84 subjects whose depressive illness went untreated from inception to resolution, the median time to recovery was 13 weeks. The cumulative monthly recovery rates were 23% after 1 month; 37% after 2 months; and 52% after 3 months. The authors concluded that "... there is a high rate of recovery in individuals not receiving somatic treatment of their depressive illness, particularly in the first 3 months of an episode."

Thus, we find no logical or empirical basis for Dr. First's assertion that, in the Clayton studies, " was common for [subjects] to develop symptoms that resembled those of a depressive episode..." (our italics). We would suggest, on the contrary, that the many manifestly depressed and bereaved subjects in Clayton's studies had developed bona fide, but rapidly resolving, MDEs in the context of bereavement, however, it is also important to note that many developed chronic and/or prolonged episodes. Incidentally, notwithstanding the epidemiological data cited by Dr. First, we simply have no way of determining what percentage of the Clayton cohort actually met modern criteria for "significant distress or impairment in one or more important areas of functioning" since this was not assessed.

Similarly, with respect to the Zisook and Shuchter study cited by Dr. First[24]: the finding that 8.6% of the 350 widows and widowers met full symptomatic criteria for an MDE immediately after the death of their spouse but were no longer depressed by the end of 2 months does not necessarily mark these as "clear-cut cases of uncomplicated bereavement." Rather, such rapidly resolving cases more likely represent instances of bereavement-related major depression that resembled the 37% of MDD patients described in the Posternak and colleagues study (ie, those whose MDD episodes resolve within 2 months. Thus, we find no basis for Dr. First's assertion that the 8.6% of cases in the Zisook-Shuchter study would have been rendered "false positives" for MDD but for the saving grace of the bereavement exclusion.

It is also important to emphasize that just as many individuals with nonbereavement-related MDEs as with bereavement-related episodes have brief depressions. Moreover, approximately equal proportions of those with bereavement-related and nonbereavement-related depression have prolonged and/or recurrent "conditions."[29,30,31] Again, these findings tell us that there is nothing special or unique about bereavement-related depression with respect to time course.

Moreover, we do not believe, to use Dr. First's phrase, that there is a substantial "...false-positive problem stemming from the symptomatic resemblance between normal grief and mild depression" (our italics). As we have argued elsewhere, there are numerous differentiating features that help distinguish "normal grief" from major depression, mild or otherwise,[32] and we have urged the incorporation of such phenomenological differences into the text of the DSM-V. To reiterate: Grief is grief; depression is depression.

Finally, we are not persuaded by Dr. First's prediction (referencing our own acknowledgment) that rates of underdiagnosis or misdiagnosis of MDD – a reality affirmed by numerous studies -- will not be greatly altered by removing the bereavement exclusion, as we advocate. We still regard the exclusion as a scientific error. While its creation was based on the groundbreaking work of Paula Clayton and colleagues, the preponderance of data since those studies were published suggest that another revision is in order. We see no virtue in enshrining the exclusion in the next DSM, however flawed the MDD criteria themselves may be. Whereas Dr. First would rather live with fewer false positives but (potentially) more false negatives, we believe that the safety and well being of bereaved patients meeting symptom and duration criteria for MDD is better served by minimizing the frequency of false negatives (ie, by erring on the side of caution and placing these patients under the rubric of MDD).

Unless and until we see convincing prospective, clinical studies, or other kinds of data(eg, reliable biomarkers, genetic and neuroendocrine studies, etc) showing that bereavement-related major depressive states differ significantly from nonbereavement-related major depression, we continue to find no clinical or scientific basis for preserving the bereavement exclusion. In closing this stimulating dialogue, we acknowledge that the definitive study capable of resolving these controversies has yet to be done. That would entail a prospective design, using reliable diagnostic instruments and outcome measures, in which subjects with bereavement-related depression (shortly after the death of a loved one) are compared with nonbereaved depressed patients, matched for severity, symptom constellation, and other relevant variables. Until such studies are done, we are left with our differing perspectives. Nonetheless, we very much appreciate Dr. First's willingness to engage with us in this important debate, and we thank Dr. Stetka for providing this forum.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.