and Quinlan, 1981; Ionescu and Popescu, 1989; Pfohl, Stangl, and Zimmerman, 1984; Shea, Glass, Pilkonis, et al., 1987). Those with personality diagnoses had an earlier age of onset for their first depressive episode, more severe depressive symptoms, more frequent episodes, longer depressive episodes, poorer short-term recovery with both antidepressant medications and psychotherapy, and more residual symptomatology at later followup. A concurrent personality disorder diagnosis is associated with a more complicated, disturbed social history, especially in patients with Cluster A or B personality disorders. When depression is complicated by personality disorder, most studies find increased rates of suicide attempts and self-harm.

The generally negative effect of personality disorders on the outcome of depression seems largely accounted for by Cluster B, or borderline personality disorder specifically. Depressed patients with co-morbid borderline personality disorder had poorer social outcomes and higher levels of residual symptomatology at both 4- and 7-year followups (Pope, Jonas, Hudson, et al., 1983). Borderline personality disorder seems highly prevalent among depressed psychiatric patients; a sample drawn from a general psychiatric population yielded an estimate of 6 percent. The estimated community prevalence of borderline personality disorder is 0.2 percent (Weissman and Myers, 1980). One study of people who were not psychiatric patients found a 1.6 percent prevalence of borderline personality disorder, but this population included a large sample of firstdegree relatives of psychiatric inpatients (Zimmerman and Coryell, 1989). There are insufficient studies of the incidence of personality disorders in primary care settings to make prevalence estimates.

Grief and Adjustment Reactions

Guideline: DSM-III-R indicates that, if depressive symptoms begin within 2 to 3 weeks of a loved one's death, the diagnosis is uncomplicated bereavement, which is not viewed as a disorder but as a normal, relatively benign state that resolves spontaneously without treatment. While uncomplicated bereavement and major depressive episodes share many symptoms, active suicidal thoughts, psychotic symptoms, and profound guilt are rare in bereavement. However, if a major depressive episode is still present 2 months following the loss, the episode is likely to be prolonged and associated with substantial morbidity. Clinically, the diagnosis of major depressive disorder may be made during the period of "grief" in those who meet the criteria for a major depressive episode 2 months following the loss. (Strength of Evidence = A.)

Grief reactions are equally common in women and men. Over the course of the first year of bereavement, Clayton (1974) found that 35 percent of 109 widows and widowers met the criteria for a major

depressive episode at 1 month, 25 percent at 7 months, 17 percent at 13 months, and 46 percent were depressed some time during the first year following the loss of a spouse.

Zisook and Shuchter (1991) studied 350 widows and widowers 2 and 13 months after their spouses' deaths. Twenty-four percent met the DSM-III-R criteria for a major depressive episode at 2 months. Since the deaths had occurred only recently, the condition of these subjects was diagnosed as uncomplicated bereavement rather than major depressive disorder. Those meeting the criteria for major depressive disorder early in the period of grief were more likely to have personal or family histories of major depressive episodes (not in response to the death of a loved one), current treatment with antidepressant medications, suicidal ideation, poor health, and poor current job satisfaction. Most important, they were likely to be in a major depressive episode 1 year following the loss. Given the anticipated prolonged suffering and disability, it is logical to consider treating these patients for the major depressive disorder, though randomized controlled trials of any treatment are lacking in this population. A distinct, but poorly studied, segment of primary care patients present serious mood symptoms thought to require clinical management, though they do not fulfill the criteria for major depressive or dysthymic disorders. DSM-III-R classifies such "subthreshold" pathology as either DNOS or an adjustment disorder with depressed mood. Little is known of the presenting symptoms, clinical course, or outcome of an adjustment disorder with depressed mood. However, the clinical utility of the diagnosis is that it allows practitioners to identify mildly distressed patients, some of whom may need followup to determine whether the symptoms remit or whether they evolve into a formal mood syndrome. If such patients develop a major depressive or dysthymic disorder, treatment should be as for the primary mood disorder. There are no randomized controlled trials of treatment for adjustment disorder with depressed mood, but it is logical to consider treatment with psychotherapy or medication for those with substantial pain, suffering, incapacity, or chronicity.

Guideline:

4 Depression Co-Occurring with

Other General Medical Disorders

Guideline: Many general medical conditions are risk factors for major depression. Major depressive disorder, when present, should be viewed as an independent condition and specifically treated. Treatment may include (a) optimizing the treatment of the general medical disorder and/or (b) providing specific treatment for the depression. (Strength of Evidence

A.)

Clinically significant depressive symptoms are detectable in

approximately 12 to 36 percent of patients with another nonpsychiatric, general medical condition. Rates in patients with specific medical disorders may be even higher. These figures far exceed the approximate 4 percent prevalence of diagnosable depression in large community samples. On the other hand, most patients with a general medical condition do not have a mood disorder. Therefore, the mood disorder, when present, should be viewed as an independent condition (perhaps precipitated by the biologic or psychological vulnerability of the individual) that should be specifically treated.

Since every co-occurrence of major depression and every general medical disorder cannot be covered in this guideline, the panel has chosen several specific examples to outline a general approach to the diagnosis of depression in patients with other medical disorders and to illustrate the primary treatment principles. Somatic symptoms are part of the syndrome of major depression, according to DSM-III-R. Many other medical disorders also cause some criterion symptoms of depression, such as weight loss, sleep disturbances, and low energy. These disorders include endocrinopathies, such as diabetes; pituitary, adrenal, or thyroid disorders; certain malignancies; some infections; some neurologic disorders; collagen disorders; cardiovascular disease; and vitamin/mineral deficiency and/or excess states. The clinician can substitute cognitive and emotional symptoms, such as fearful or depressed appearance, social withdrawal or decreased talkativeness, brooding, self-pity, or pessimism and unreactive mood for the standard DSM-III-R somatic symptoms when there is concern that the suspected concurrent medical disorder may be causing the criterion somatic symptoms of depression (Endicott, 1984; Kathol, Mutgi, Williams, et al., 1990).

Once the syndrome of depression has been identified in patients with a general medical illness, the differential causes of depressive

symptomatology must be reviewed to make sure the appropriate treatment is administered. The risk factors associated with primary mood disorders

should be reviewed to determine whether the patient's condition fits a typical picture of primary mood disorder or whether alternative causes can explain the depressive syndrome or symptoms.

When depression and another medical condition occur together, there

are several logically plausible explanations:

■ The general medical disorder biologically causes depression; for example, hypothyroidism may cause depressive symptoms.

■ The general medical disorder triggers the onset of the depression in those who are genetically vulnerable to depressive disorders; for example, Cushing's disease may precipitate a major depressive episode. ■ The general medical disorder psychologically causes the depression; for example, a patient with cancer may become clinically depressed as a psychological reaction to the prognosis, pain, and incapacity.

■ The general medical disorder and the mood disorder are not causally related.

It is important for the practitioner to differentiate among these options for patients with depressive and other psychiatric or medical conditions. In the first two instances, treatment aims first at the general medical disorder. If the depression persists, it is treated once the general medical disorder is stabilized. In the third case, the general medical disorder is treated while counseling, education, support, and medication are used to treat the depression. In the last instance, specific treatment is initiated for both disorders (Figure 5; see also Figure 2). While one uncontrolled study (Hall, Gardner, Stickney, et al., 1980) suggests that depressive symptoms resolve with treatment of the general medical illness alone in more than 60 percent of patients with depression associated with treatable general medical disorders, the prognosis for such patients remains ill-defined.

Once it has been established that the depressive symptoms are due to a primary mood disorder, treatment is aimed at the mood disorder. If, on the other hand, the depression is caused by the general medical condition, several additional steps are necessary. First, treatment of the general medical illness should be optimized. Thereafter, sufficient time should be allowed for this treatment to alter the course of the mood symptoms. If the patient's mood disorder or symptoms do not respond to treatment for the general medical illness, or if the patient has an illness, such as cancer or diabetes, that is under optimal control but is not curable, the depression should be treated as a primary mood disorder.

Stroke

Guideline: Depression following stroke is not fully explained as a psychological response to the associated impairment. There appear to be subgroups of depressed post-stroke patients whose depression is causally related to the injury, possibly including its strategic location in the brain (left dorsal lateral frontal cortex or left basal ganglia);

Figure 5. Relationship between major depressive and other current general medical disorders

Note: In some clinical situations, treatment of the depression (e.g., if severe, incapacitating, or life-threatening) cannot be delayed until treatment for the general medical disorder has been optimized.

a family history of depression; premorbid subcortical atrophy; and premorbid or ongoing social factors. When a patient with a recent stroke meets the criteria for a major depressive episode, organic (secondary) mood disorder is diagnosed. (Strength of Evidence = B.)

The association between cerebral infarction and depression has long been recognized. However, systematic studies (Depression Guideline Panel, forthcoming) have found only a weak relationship between depression severity and physical/cognitive impairment following stroke. Case reports (Ross and Rush, 1981) indicate that post-stroke patients who are also depressed, especially those with major depressive disorder, are less