who developed major depressive disorder had experienced at least one prior major depressive episode that was separate from the onset of panic attacks. In a comparative family study of individuals with panic disorder with agoraphobia, limited phobic avoidance, or social phobia, a significantly higher percentage of agoraphobic patients had family histories of mood disorders compared to those with limited and social phobias (Munjack and Moss, 1981). Available data are consistent with the idea that many persons with concurrent major depressive disorder and panic, social phobic, or generalized anxiety disorders may actually have only a single disorder that presents with both anxiety and depressive symptoms. The decision about which disorder to treat may in some cases be determined by one of the following: ■ The patient's family history. The symptom complex that began first in this episode of illness. Given that followup studies of those with panic or other anxiety disorders reveal that many will subsequently develop major depressive disorder and that they often have had a prior major depressive disorder or have a family history of major depressive disorder, the depression is the appropriate main target of treatment in many cases. Sometimes, however, only a treatment trial and observation will answer this complex diagnostic question. Whichever disorder is primary, the data are clear that the combination of panic and major depressive disorders results in a more severe disorder with greater impairment than does either disorder alone. For example, depressed patients with associated panic attacks have a more severe depressive illness and are less likely to recover during a 2-year followup than are those without panic attacks (Coryell, Endicott, Andreasen, et al., 1988). The lifetime suicide attempt rate for persons with both panic and major depressive disorders is more than twice that of those with panic disorder, but without major depressive disorder (19.5 versus 7.0 per 100) (Johnson, Weissman, and Klerman, 1990). In two separate studies, panic disorder and primary major depressive disorder were each associated with high suicide rates (Coryell, Noyes, and Clancy, 1982, 1983). These data strongly suggest the importance of inquiring about, and even expecting to find, a concurrent mood disorder (especially major depressive disorder) in patients with anxiety complaints. If an individual presents with both conditions and if they are equally impairing, the practitioner should consider treatment with medications for which efficacy has been demonstrated for both conditions. These include MAOIS, SSRIs, TCAs, or in selected cases alprazolam. Eating Disorders Guideline: The practitioner is advised to ask about anorexia nervosa and bulimia nervosa in young women who present with any mood disorder, especially those with amenorrhea. If present, the eating disorder is the principal target of treatment. (Strength of Evidence = B.) The eating disorders include anorexia nervosa and bulimia nervosa (American Psychiatric Association, 1987). Anorexia is a refusal to maintain body weight over a minimal normal weight, accompanied by intense fear of becoming fat, disturbance in body image, and amenorrhea. Anorexia occurs in 0.2 to 0.8 percent of adolescent girls in school cohort studies and in 0.05 to 0.1 percent of adults in community samples (Robins, Helzer, Weissman, et al., 1984). Bulimia is characterized by recurrent episodes of rapid consumption of large amounts of food, accompanied by a feeling of loss of control; regular use of vomiting, laxatives, or other means to attempt to control weight; and overconcern with body shape and weight. The prevalence of bulimia is approximately 1 percent among adolescent and young adult women. Prevalence rates of 1.8 to 1.9 percent have been found in populations at family planning clinics, and rates ranging from 1 to 22 percent have been found in primary care settings. No large studies of the prevalence of eating disorders among patients with major depressive disorder have been conducted. On the other hand, it is well established that one-third to one-half of patients with eating disorders (either anorexia or bulimia) suffer concurrently from a major depressive syndrome. Approximately 50 to 75 percent of eating disorder patients have a lifetime history of major depressive disorder. Dysthymic disorder and DNOS occur less frequently among patients with eating disorders by most reports. Assuming a 1 percent prevalence of eating disorders and an 8 percent prevalence of major depressive disorder, and assuming that one-half of eating disorder patients also suffer from concurrent major depressive disorder, the likelihood of an eating disorder may be as high as 1/16 (6 percent) in women between the ages of 15 and 35 who suffer from a major depressive disorder. Patients with undernutrition from various etiologies often exhibit depressive symptoms, including depressed mood, irritability, poor concentration, indecisiveness, loss of sexual interest, and sleep disturbance, all of which usually improve with weight gain. Thus, when significant depressive symptoms are found with anorexia nervosa, treatment is first aimed at the eating disorder. If mood symptoms persist after the malnourished state has been reversed, treatment is as for a primary mood disorder. A number of studies indicate that some antidepressant medications (e.g., imipramine, desipramine, fluoxetine, MAOIs) and formal cognitive behavioral psychotherapies may help treat the bulimia with or without associated depressive symptoms. (For examples, see Hughes, Wells, Cunningham, et al., 1986; Mitchell and Groat, 1984; Pope, Hudson, Jonas, et al., 1983; Walsh, Stewart, Wright, et al., 1982.) Such treatments, if successful, usually result in remission of the depressive symptoms. Obsessive-Compulsive Disorders Guideline: For those depressed patients whose disorder has some obsessive features, the mood disorder is the initial focus of treatment. If full-blown OCD is present with depressive symptoms or manicdepressive disorder, the OCD is usually the initial objective of treatment. Evidence from OCD medication treatment trials suggests that, if the OCD is treated successfully, the depressive symptoms usually abate. (Strength of Evidence = A.) The ECA survey revealed some overlap between OCD, major depressive disorder, and schizophrenic disorder. Most studies agree that the lifetime occurrence of depressive symptoms is high (80 to 100 percent) among OCD patients, even though only about 10 to 30 percent of them meet the criteria for major depressive disorder at the time of admission to study. The major depressive disorder usually follows the onset of OCD, while schizophreniform symptoms are equally likely to precede or to follow OCD onset. A review of 13 followup studies found that OCD patients were at increased risk for major depressive disorder, but not for schizophrenia (Goodwin and Jamison, 1990). Whether currently or ever depressed, OCD patients are likely to have a family history of depression. The ECA data reanalysis showed that OCD is relatively common among those with a mood disorder complicated by the presence of another psychiatric condition (Johnson and Weissman, unpublished manuscript). Specifically, OCD was found in 35 percent of the 31 subjects with major depressive disorder, in 15 percent of the 46 with dysthymic disorder, and in 40 percent of the 167 with DNOS, when the mood disorder was associated with another psychiatric condition. The practitioner must differentiate severe depression, which may present with obsessive features, from true OCD. Severely depressed patients have recurrent ruminations, but rarely have compulsions. The content of their ruminations is usually consistent with a negative sad mood (e.g., guilty preoccupations). These patients often do not meet the formal criteria for OCD and often have had prior episodes of severe depression. The onset of these obsessive symptoms in the severely depressed occurs at the same time as the onset of the major depressive episode. Somatization Disorder Guideline: Somatization is defined as the presentation of somatic symptoms by patients with underlying psychiatric illness or psychosocial distress. These somatic symptoms have no, or insufficient, underlying organic cause. While most depressed patients have medically unexplained somatic complaints, they are rarely of sufficient intensity or frequency to meet the threshold for somatization disorder. (Strength of Evidence = A.) Somatization may well be the main reason for the misdiagnosis of mental illness by primary care physicians (Bridges and Goldberg, 1985). In primary care settings, many depressed and nondepressed patients present with medically unexplained symptoms. Most patients with such complaints do not meet the formal criteria for somatization disorder, which in DSM-III-R require the presence of 13 or more medically unexplained symptoms. Most patients with diffuse unexplained somatic symptoms in primary care and community samples either have a treatable psychiatric illness (e.g., anxiety or depressive disorders) or are responding to stressful life events. Accurate differential diagnosis and treatment of the acute psychiatric illness often decrease the tendency toward somatization. Twelve studies of a total of 976 primary care and medical specialty (e.g., gynecology) patients with clinically significant depressive symptoms revealed that 30 to 87 percent also had clinically significant pain complaints (Fishbain, Goldberg, Meagher, et al., 1986; Haley, Turner, and Romano, 1985; Katon, 1988; Katon, Ries, and Kleinman, 1984; Kramlinger, Swanson, and Maruta, 1983; Large, 1986; Lindsay and Wyckoff, 1981; Maruta, Vatterott, and McHardy, 1989; Schaffer, Donlon, and Bittle, 1980; Turner and Romano, 1984; Walker and Greene, 1989; Walker, Katon, Harrop-Griffiths, et al., 1988). A highly significant 30 to 50 percent met the formal criteria for major depressive disorder. Two additional studies found that 50 to 70 percent of primary care patients with psychiatric illness (mainly mood disorders) presented with somatic complaints (Katon, 1987; Katon, Kleinman, and Rosen, 1982). Conversely, pain symptoms occurred in approximately 60 percent of patients with major depressive disorder in three studies with 403 primary care and medical specialty patients, in which pain was documented by selfreport (Lindsay and Wyckoff, 1981; Magni, Schifano, and de Leo, 1985; von Knorring, Perris, Eisemann, et al., 1983). Major depressive disorder patients had significantly more symptoms on a medical review of symptoms, even when the investigators controlled for chronic medical illness. In a study using a 1,000-patient sample from a health maintenance organization, patients with one pain complaint were no more depressed than were controls (Dworkin, von Korff, and LeResche, 1990). However, patients with two pain complaints were six times more likely, and patients with three pain complaints were eight times more likely, to have a clinical depression. The Medical Outcomes Study also showed that patients with major depressive disorder perceived their general health as poorer; had more limitations in physical, social, and vocational functioning; and had more pain complaints than did persons with chronic medical illnesses (Wells, Stewart, Hays, et al., 1989). Guideline: The practitioner is advised to have a high index of suspicion for major depressive or other mood disorders if patients present with two or more unexplained pain complaints. A formal diagnostic evaluation for mood disorders is recommended. (Strength of Evidence B.) = The majority of patients with major depressive disorder, DNOS, or dysthymic disorder have some pain symptoms. The most common complaints in such patients are joint pain, headaches, backaches, and abdominal pain. Treatment of major depressive disorder with somatic complaints usually results in complete relief of the pain complaints. Personality Disorders Guideline: Personality disorders are not uncommon among mooddisordered patients. The presence of a personality disorder does not exclude diagnosis of a mood disorder, if present. When both a major depressive and a personality disorder are present, more frequent and longer major depressive episodes, as well as poorer interepisode recovery (if untreated), may be anticipated. For some with major depression, symptoms that initially appear to be maladaptive personality traits remit once the depressive disorder improves. (Strength of Evidence = B.) Studies of depressed patients using structured interviews indicate prevalences of personality disorders ranging from 35 percent in psychiatric outpatients to 72 percent in psychiatric inpatients. Most studies report rates of 45 to 65 percent. However, most of these studies were conducted in specialized research centers, which may attract a greater proportion of mood-disordered patients who also have personality disorders. The DSM-III-R personality disorders are grouped into three clusters: Cluster A (odd/eccentric)—paranoid, schizoid, and schizotypal disorders. ■ Cluster B (dramatic/emotional/erratic)-antisocial, borderline, histrionic, and narcissistic disorders. ■ Cluster C (anxious/fearful)-avoidant, dependent, obsessive-compulsive, and passive-aggressive disorders. The presence of these personality disorders may negatively affect the natural course and treatment response of mood disorders. Five studies have examined patients with concurrent major depressive disorder and a personality disorder (Black, Bell, Hulbert, et al., 1988; Charney, Nelson, |