Comorbidity and Depression Treatment K. Ranga Rama Krishnan Comorbidity is common among patients with major depression, but in most instances it may be of little relevance. Nonetheless, it is a complex issue because of its relation to treatment response, and few studies have attempted to address this. Most have examined comorbidity after the fact in secondary analyses. In this article, I focus on whether comorbidity influences depression treatment response among patients who are primarily diagnosed as suffering from major depression. At least three comorbidities are believed to influence treatment response: medical, anxiety, and personality disorders. Whether studies find that these factors predict worse outcomes in patients with major depression appears to depend on the nature and severity of the medical illness, the study setting, and the study design. The best designed studies reported the least effects of these factors on treatment outcome. Clinically, this suggests that these factors should not be seen as impediments to treatment. Biol Psychiatry 2003 Society of Biological Psychiatry Key Words: Comorbidity, medical, treatment, anxiety, response, personality Introduction Comorbidity is common among patients with major depression, but in most instances it may be of little relevance. Nonetheless, it is a complex issue because of its relation to treatment response, and few studies have attempted to address this. Most available studies examined comorbidity after the fact in secondary analyses. In this article, I focus on whether comorbidity influences depression treatment response among patients who are diagnosed primarily as suffering from major depression. It is important to keep in mind that depression secondary to other diseases may not have the same response pattern. At least three comorbidities are believed to influence treatment response: medical, anxiety, and personality disorders. I briefly review the available information on whether the presence of these three categories of comorbidity influence treatment response among patients with a primary diagnosis of major depression. From the Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, North Carolina. Address reprint requests to Ranga R. Krishnan, M.D., Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Box 3950, Durham NC 27710. Received May 3, 2002; revised August 26, 2002; accepted October 2, 2002. Medical Comorbidity Clinical trials of antidepressants generally exclude patients who have significant medical illness, yet depression with medical comorbidity is the norm rather than the exception among patients who are seen in most clinical settings. The treatment of depression in medically ill patients is challenging. Recognition, compliance, differential diagnosis, side effects, and tolerance of drug regimens can complicate the treatment of depression among patients who are medically ill. Before the introduction of selective serotonin reuptake inhibitors (SSRIs), it was generally believed that medically ill depressed patients did not tolerate or respond well to antidepressant treatment. Koenig et al (1989) reported that nortriptyline was contraindicated in 90% of medically ill depressed patients. He also noted that 80% of potentially eligible patients were unable to complete a trial of nortriptyline. In a retrospective review Popkin et al (1985) reported that only 40% of medically ill depressed patients responded to treatment and that 32% could not tolerate treatment. Both these studies were conducted before the advent of SSRIs and led to a nihilistic perception regarding the use of antidepressants in the medically ill. This perception may not be warranted. A recent Cochrane report on the use of antidepressants in the medically ill addresses the issue of whether antidepressants are effective in this population (Gill and Hatcher 2000). This review analyzed all relevant randomized trials that compared any antidepressant drug with placebo or no treatment in patients diagnosed with depression and a specified physical disorder. This review included 18 studies, covering 838 patients with a range of medical conditions (cancer in two studies, diabetes in one, head injury in one, heart disease in one, HIV in five, lung disease in one, multiple sclerosis in one, renal disease in one, stroke in three, combined disorders in two). Six studies used SSRIs, three used atypical antidepressants, and the reminder used tricyclic antidepressants (TCA). The key finding was that patients treated with antidepressants were more likely to improve than those who were given placebo or no treatment. The finding that about four patients would need to be treated with antidepressants to produce one recovery from depression that would not have occurred had they been given placebo or no treatment is similar to that seen in trials of depressed patients without medical problems. The other interesting finding was that antidepressants were well tolerated by patients; about 10 patients would need to 2003 Society of Biological Psychiatry 0006-3223/03/$30.00 doi:10.1016/s0006-3223(03)01787-0
702 BIOL PSYCHIATRY K.R.R. Krishnan be treated with antidepressants to produce one dropout from treatment, which would not have occurred had they been given placebo. (Gill and Hatcher 2002). By inference from this review, one can conclude that medical comorbidity is not a major factor in treatment response. A problem with the studies reviewed in Gill and Hatcher (2002) is that they did not include both healthy and medically ill patients with depression. Thus, it cannot be definitively stated that, under the same trial conditions, medical comorbidity does not influence treatment response. Steffens recently studied this issue in a naturalistic manner (Steffens et al 2002). Thirty-one elderly patients with unipolar major depression (DSM-IV) who were enrolled in Duke s Mental Health Clinical Research Center for the Study of Depression in Later Life were prescribed bupropion SR or IR, alone or in combination with other antidepressant agents, for 12 weeks. Montgomery Asberg Depression Rating Scale (MADRS; Montgomery and Asberg 1979) scores and Clinical Global Impression (CGI) severity scores were used to define response. Seventy-four percent of the sample responded to treatment. Fifty-three percent (16 of 30) achieved a partial or complete remission of major depression at week 12. Response rates did not differ between those with high versus low medical comorbidity. Steffens et al (2002) concluded that geriatric patients with high and low medical comorbidity responded well to bupropion and bupropion SR. We recently evaluated the effect of sertraline in patients with major depression and comorbid vascular disease (Krishnan et al 2001). Patients were divided into one of three groups: 1) patients with a current diagnosis of hypertension only, 2) patients with a current or past history of cardiovascular illness but no hypertension, and 3) patients with no hypertension or comorbid vascular illness. Sertraline treatment yielded similar levels of response in all three groups (response criterion: CGI much or very much improved) at treatment end point on a completer analysis (hypertension: 86%; vascular disease: 89%; no vascular disease: 77%). Both our study and that of Steffens et al (2002) were limited to a single drug treatment, however, and the medical comorbidity evaluated was restricted to vascular disease in our study. We therefore evaluated the role of medical factors in a large sample of elderly patients who were treated using a staged approach. The sample consisted of 259 subjects enrolled in the National Institute of Mental Health sponsored Mental Health Clinical Research Center for the Study of Depression in Later Life (Conte Center). The subjects were 60 years and met DSM-IV criteria for major depression at baseline. Exclusions included other major psychiatric illness (e.g., schizophrenia, schizoaffective disorder, bipolar disorder) or major neurologic illness (e.g., Alzheimer s Figure 1. Cumulative remission for depressed patients starting from baseline. disease and other forms of dementia, stroke, Parkinson s disease, and multiple sclerosis); subjects with current or recent histories of substance abuse were also excluded. At baseline, subjects received standardized clinical assessments including the MADRS and Cumulative Illness Rating Scale (CIRS; Linn et al 1968). A trained interviewer administered the Duke Depression Evaluation Schedule (DDES) (George et al 1989; Landerman et al 1989; Robins et al 1981). The DDES is a composite instrument that includes the Centers for Epidemiologic Studies-Depression Scale (CES-D) (Radloff 1977) and portions of the National Institute of Mental Health Diagnostic Interview Schedule (DIS). The interview consists of questions that screen for DSM-IV diagnoses including major depression, bipolar disorder, generalized anxiety disorder, and panic disorder. Clinical assessments were repeated every 3 months and when contact was clinically indicated. Subjects in the study were treated as clinically indicated, using antidepressant medications, electroconvulsive therapy (ECT), and individual and group cognitive behavioral psychotherapy as described in Steffens et al (2001). The results of the study demonstrated that the cumulative remission rate was slightly greater among patients without medical illness than among those with mild medical illness (CIRS 5; see Figure 1). Remission was defined as MADRS score less than 8; when adjusted for age, however, this variable was not significant. In summary, the results of these three studies, in consonant to the Cochrane review (Gill and Hatcher 2002), indicate that medical comorbidity has only a modest and insignificant effect on antidepressant treatment response; however, this may not be the case for other medical conditions and needs to be further evaluated. Whether studies find that medical disorders predict worse outcome in patients with major depression appears to depend on the nature and severity of
Comorbidity and Depression Treatment BIOL PSYCHIATRY 703 the medical illness, the study setting, and the study design. The best designed studies reported the least effect of medical illness on depression treatment outcome. Clinically, this suggests that medical illness should not be seen as an impediment to treatment. Anxiety Disorder Anxiety disorder is also common among patients with major depression. Zimmerman (2002) studied the frequency of diagnostic comorbidity in major depressive disorder (MDD). At the time of the evaluation, 64.1% of the patients met criteria for at least 1 of the 23 Axis I disorders, and more than one third had two or more disorders. Anxiety disorders were the most frequent comorbid disorders (56.8%), and social phobia was the most frequent individual disorder. Among depressed patients, the more severe the depression, the more likely the presence of anxiety symptoms. It is assumed that anxiety symptoms affect treatment response. This has been evaluated in a number of studies. Paykel (1972) reported that people with depression who were anxious responded poorly to amitriptyline compared with other depressed patients. Another study reported that neurotic symptoms also predicted poor response to TCAs (Bielski 1976). The data with regard to the SSRIs appears to be different. Tollefson et al (1994) pooled data from 19 randomized, double-blind clinical trials comparing fluoxetine with placebo or TCAs in patients with major depression. On the basis of the anxiety/somatization factor within the 21-item Hamilton Rating Scale for Depression (HAM- D), patients were characterized as anxious (score 7) or nonanxious (score 7). Fluoxetine was significantly (p.05) more effective than placebo in treating both anxious and nonanxious major depression. Fluoxetine was also more effective than placebo in reducing the anxiety/ somatization factor score (Tollefson et al 1994). The effect of anxiety on treatment response to sertraline was evaluated in a study of chronic depression (Russell et al 2001). In this study, patients diagnosed with chronic major or double depression were randomized to 12 weeks of double-blind treatment with either sertraline or imipramine in a 2:1 ratio. A high-anxiety subgroup was operationally defined by a HAM-D anxiety/somatization factor score 7. Of the total sample, 209 were treated with imipramine and 426 with sertraline. Thirty-six percent of the population met criteria for the high-anxiety subgroup. Patients with significant concurrent anxiety symptoms were more likely to respond by 12 weeks (66.4%) than those without significant anxiety symptoms (54.2%). There was no significant difference in response rates for sertraline versus imipramine. In a comparison of SSRIs among patients with anxious depression classified as described earlier, no difference in treatment response was found among paroxetine, sertraline, and fluoxetine (Fava et al 2000). Rudolph (1998) evaluated the efficacy of venlafaxine in patients with anxiety and depression. A pooled analysis was conducted of six short-term trials of venlafaxine measuring anxiety in anxious-depressed patients using the HAM-D, anxiety/somatization factor and psychic anxiety item scores. Treatment with venlafaxine resulted in a significant improvement in depression scores in patients who were anxious at baseline compared with placebo-treated patients. Data from eight randomized, double-blind, placebocontrolled clinical trials comparing mirtazapine to placebo in patients with high anxiety and depression were reported by (Fawcett et al 1998). Mirtazapine-treated patients demonstrated a statistically significant reduction in the sum of anxiety and agitation compared with placebotreated patients. In general, the drug was effective in treating both anxiety and depression, and anxiety did not predict treatment response (Fawcett et al 1998). All these studies primarily classified patients on the basis of the anxiety subscale of the HAM-D. They did not specifically address the issue of whether the anxiety was part of depression or as part of another anxiety disorder. In general, these studies would have excluded patients with primary anxiety disorders. The data are modest in terms of whether the presence of an anxiety disorder can influence treatment response. Sonawalla (1999) assessed treatment response in patients with major depression and comorbid anxiety disorder who were treated with fluvoxamine. The mean number of comorbid anxiety disorders per patient was 2.1 1.1. Fluvoxamine was shown to be effective in treating outpatients with major depression with comorbid anxiety disorder. This study was limited by the small number of subjects and by the fact that a number of studies have not shown a clear antidepressant effect for fluvoxamine. Silverstone et al (2001) recently reported on the efficacy of venalfaxine on patients with concomitant generalized anxiety disorder and depression. Ninety-two patients meeting DSM-IV criteria for MDD who also had comorbid generalized anxiety disorder were compared with 276 noncomorbid patients. Patients received venlafaxine XR (75 225 mg/day), fluoxetine (20 60 mg/ day), or placebo for 12 weeks. Onset of efficacy was slower in comorbid than in noncomorbid patients. Response, defined as 50% decrease in symptom score on the HAM-D, was achieved in 66% of the comorbid patients in the venlafaxine XR group at week 12. This response was higher than that seen with fluoxetine (52%) or placebo (36%). In contrast to these reports, a study in primary care patients reported that anxiety can predict persistence of depression (Gaynes et al 1999). Of 85 patients with major
704 BIOL PSYCHIATRY K.R.R. Krishnan depression at baseline, 43 had coexisting anxiety disorder (38 with social phobia). The risk for persistent depression at 12 months was 44% greater among those with coexisting anxiety; this may largely reflect the type of anxiety disorder that coexists with major depression. A recent study (Hoehn-Saric et al 2000) showed that sertraline is more effective than desipramine for depression in the context of obsessive compulsive disorder. This is consistent with data showing that SSRIs are more effective than drugs that primarily work through a norepinephrine mechanism. Given that SSRIs are effective for posttraumatic stress disorder, social phobia, obsessive compulsive disorder, panic disorder, and generalized anxiety disorder, it is not surprising that these drugs are effective in treating depression in the context of anxiety disorders. Nonetheless, randomized trials evaluating whether the response rate is altered by the presence of each of these anxiety disorders remain sparse. The data with psychotherapy may be significantly different. For example, Feske et al (1998) reported that higher levels of anxiety predicted poor response to treatment. Additional studies are needed to evaluate whether treatment response is different with specific comorbid anxiety disorder such as obsessive compulsive disorder and social phobia. Personality Traits and Disorders Personality traits have long been considered important in the treatment outcome of depression. The study of personality traits and disorders is complex, and a variety of methods has been used, including dimensional measures, such as neuroticism and the Tridimensional Personality Questionnaire (TPQ) (Newman et al 2000), and categorical measures such as DSM-based personality disorder scales. Mulder (2002) recently evaluated these studies in an excellent review. To summarize the results based on his review, neuroticism generally predicted poor long-term response. The studies of shorter duration care were more equivocal, with most not showing a relationship to shortterm outcome. Studies with the TPQ were done at a later time point and were better designed. The TPQ measures novelty seeking, which reflects differences in the behavioral activation system; harm avoidance, which reflects differences in the behavioral inhibition system; and reward dependence, which reflects differences in the behavioral maintenance system. The largest study, by Nelson and Clonninger (1997), showed that reward dependence predicted less than 1% of variance. The categorical assessment studies were highly varied in that a variety of methods was used. Seven studies used both a standardized assessment and standardized treatment (sertraline, TCAs, ECT). Of the five studies that used Table 1. Personality Disorder and Treatment Response Study N Treatment Results Sato (1993) 96 Maprotiline Those with personality disorder less like to recover Fava (1994) 83 Fluoxetine No difference for cluster A or C; cluster B better outcome Joyce (1994) 84 Clomipramine/ No overall difference desipramine Hirshfeld (1998) 623 Sertraline/ Not a significant factor Imipramine Peselow (1992) 44 Desipramine More personality disorder poorer response medication treatment, three showed no difference, and two showed worse outcome for patients with personality disorders (one with desipramine and one with maprotoline; Table 1). The largest study (Hirshfield 1998), with more than 600 patients, used sertraline and showed no difference between patients with and without significant personality pathology. Fava (2000) reported that cluster B factors affected treatment response, with patients with high reward dependence responding better to clomipramine than to desipramine. In his review of these studies, Mulder s appropriate conclusion is as follows: Whether or not personality pathology significantly worsens outcome in patients with major depression appears to depend on study design, since the rate of personality pathology varies markedly depending on how it is measured. In addition, depressed patients with personality pathology appear less likely to receive adequate treatment in uncontrolled studies. Finally, studies rarely control for depression characteristics (e.g., chronicity, severity) that may influence outcome and be related to personality pathology. Overall, the best designed studies reported the least effect of personality pathology on depression treatment outcome. Clinically, this suggests that comorbid personality pathology should not be seen as an impediment to good treatment response (Mulder 2002). Summary Comorbidity as generally seen in the context of major depression appears to be only a minor factor determining treatment response. Medical comorbidity has a modest and insignificant effect on short-term treatment response to SSRIs and other, more recently introduced medications. The nihilistic assumption that antidepressants are not effective based on early studies with TCAs is not warranted. In fact, even in patients with significant medical problems, antidepressants are effective. Severity of anxiety symptoms does not appear to have a major differential
Comorbidity and Depression Treatment BIOL PSYCHIATRY 705 effect on somatic treatment response, especially for the SSRIs and other newer antidepressants, although it may for psychotherapy. Anxiety disorders are well treated with SSRIs, and thus the findings that these drugs work well for depression in the context of anxiety disorders is not surprising. At least for comorbid generalized anxiety and depression, the presence of comorbidity does not affect treatment response. Personality disorder also does not appear to have a major effect on treatment response. Although intuitively it is likely that severe personality disorder is likely to reduce treatment response, this has yet to be studied. Neuroticism appears to be a predictor of long-term treatment response. Aspects of this work were presented at the conference, Difficult-to- Treat Depression, held April 21 22, 2002 in San Francisco, California. The conference was sponsored by the Society of Biological Psychiatry through an unrestricted grant provided by Eli Lilly and Company. References Bielski RJ, Friedel RO (1976): Prediction of tricyclic antidepressant response: A critical review. Arch Gen Psychiatry 33:1479 1489. Fava M, Bouffides E, Pava JA, McCarthy MK, Steingard RJ, Rosenbaum JF (1994): Personality disorder comorbidity with major depression and response to fluoxetine treatment. Psychother Psychosom 62:160 167. Fava M, Rosenbaum JF, Hoog SL, Tepner RG, Koop JB, Nilsson ME (2000): Fluoxetine versus sertraline and paroxetine in major depression: Tolerability and efficacy in anxious depression. Affect Disord 59:119 126. Fawcett J, Barkin RL (1998): A meta-analysis of eight randomized, double-blind, controlled clinical trials of mirtazapine for the treatment of patients with major depression and symptoms of anxiety. J Clin Psychiatry 59:123 127. Feske U, Frank E, Kupfer DJ, Shear MK, Weaver E (1998): Anxiety as a predictor of response to interpersonal psychotherapy for recurrent major depression: An exploratory investigation. Depress Anxiety 8:135 141. Gaynes BN, Magruder KM, Burns BJ, Wagner HR, Yarnall KS, Broadhead WE (1999): Does a coexisting anxiety disorder predict persistance of depressive illness in primary care patients with major depression? Gen Hosp Psychiatry 21(3): 158 167. George LK, Blazer DG, Hughes DC, Fowler N (1989): Social support and the outcome of major depression. Br J Psychiatry 154:478 485. Gill D, Hatcher S (2000): Antidepressants for depression in medical illness. Cochrane Database Syst Rev. (4):CD001312. Hirschfeld RM, Russell JM, Delgado PL, Fawcett J, Friedman RA, Harrison WM, et al (1998): Predictors of response to acute treatment of chronic and double depression with sertraline or imipramine. J Clin Psychiatry 59:669 675. Hoehn-Saric R, Ninan P, Black DW, Stahl S, Greist JH, Lydiard B, et al (2000): Multicenter double-blind comparison of sertraline and desipramine for concurrent obsessive-compulsive and major depressive disorders. Arch Gen Psychiatry 57:76 82. Joyce PR, Mulder RT, Cloninger CR (1994): Temperament predicts clomipramine and desipramine response in major depression. J Affect Disord 30:35 46. Koenig HG, Goli V, Shelp F, Kudler HS, Cohen HJ, Meador KG, Blazer DG (1989): Antidepressant use in elderly medical inpatients: Lesions from an attempted clinical trial. J Gen Intern Med 4:498 505. Krishnan KR, Doraiswamy PM, Clary CM (2001): Clinical and treatment response characteristics of late-life depression associated with vascular disease: A pooled analysis of two multicenter trials with sertraline. Prog Neuropsychopharmacol Biol Psychiatry 25:347 361. Landerman R, George LK, Campbell RT, Blazer DG (1989): Alternative models of the stress buffering hypothesis. Am J Community Psychol 17:626 642. Linn BS, Linn MW, Gurel L (1968): Cumulative illness rating scale. J Am Geriatr Soc 16:622 626. Montgomery SA, Asberg M (1979): A new rating scale designed to be sensitive to change. Br J Psychiatry 134:382 389. Mulder RT (2002): Personality pathology and treatment outcome in major depression: A review. Am J Psychiatry 159:359 371. Nelson E, Cloninger CR (1997): Exploring the TPQ as a possible predictor of antidepressant response to nefzaodone in a large multi-site study. J Affect Dis 44:197 200. Newman JR, Ewing SE, McColl RD, Borus JS, Nierenberg AA, Pava J, Fava M (2000): Tridemensional Personality Questionnaire and treatment response in major depressive disorder: a negative study. J Affect Dis 57:241 247. Paykel ES (1972): Depressive typologies and response to amitriptyline. Br J Psychiatry 120:147 156. Peselow ED, Fieve RR, DiFiglia C (1992): Personality traits and response to desipramine. J Affect Disord 24:209 216. Popkin MK, Callies AL, Mackenzie TB (1985): The outcome of antidepressant use in the medically ill. Arch Gen Psychiatry 42:1160 1163. Radloff LS (1977): The CES-D scale: A self-report depression scale for research in the general population. Appl Psychol Meas 1:385 401. Robins LN, Helzer JE, Croughan J, Ratcliff KS (1981): National Institute of Mental Health Diagnostic Interview Schedule: Its history, characteristics, and validity. Arch Gen Psychiatry 38:381 389. Rudolph RL, Entsuah R, Chitra R (1998): A meta-analysis of the effects of venlafaxine on anxiety associated with depression. J Clin Psychopharmacol 18:136 144. Russell JM, Koran LM, Rush J, Hirschfeld RM, Harrison W, Friedman ES, et al (2001): Effect of concurrent anxiety on response to sertraline and imipramine in patients with chronic depression. Depress Anxiety 13:18 27. Sato T, Sakado K, Sato S (1993): Is there any specific personality disorder or personality disorder cluster that worsens the short-term treatment outcome of major depression? Acta Psychiatr Scand 88:342 349.
706 BIOL PSYCHIATRY K.R.R. Krishnan Silverstone PH, Salinas E (2001): Efficacy of venlafaxine extended release in patients with major depressive disorder and comorbid generalized anxiety disorder. J Clin Psychiatry 62:523 529. Sonawalla SB, Spillmann MK, Kolsky AR, Alpert JE, Nierenberg AA, Rosenbaum JF, Fava M (1999): Efficacy of fluvoxamine in the treatment of major depression with comorbid anxiety disorders. J Clin Psychiatry 60(9):580 583. Steffens DC, Doraiswamy PM, McQuoid DR (2001): Bupropion SR in the naturalistic treatment of elderly patients with major depression. Int J Geriatr Psychiatry 16:862 865. Steffens DC, McQuoid DR, Krishnan KRR (2002): The Duke somatic treatment algorithm for geriatric depression (STAGED) approach. Psychopharm Bull 36:58 68. Tollefson GD, Holman SL, Sayler ME, Potvin JH (1994): Fluoxetine, placebo, and tricyclic antidepressants in major depression with and without anxious features. J Clin Psychiatry 55:50 59. Zimmerman M, Chelminski I, McDermut W (2002): Major depressive disorder and Axis I diagnostic comorbidity. J Clin Psychiatry 63:187 193.