Current Perspectives on Management of Co-Morbid Depression in COPD

Abstract

Individuals with COPD have a higher prevalence of co-morbid depression than either the general population or patients with other chronic illnesses. The best estimates report a prevalence of approximately 40% in COPD patients, compared to 15% in the general population. Depression in COPD patients leads to a lower quality of life, greater objective impairment in function, and decreased adherence to therapeutic interventions. While many depressed COPD patients have been treated empirically with antidepressants—subjecting them to antidepressant side effects, toxicities, and costs—there is a surprising lack of evidence supporting or directing that treatment. We review the current literature regarding the management of depression in COPD, suggest strategies for management, and future research needs.

• Chronic obstructive pulmonary disease
• Depression
• Antidepressants

Introduction

COPD is the fourth leading cause of death in the United States afflicting up to 24 million individuals in the United States alone [[1]]. While deaths attributable to cardiovascular diseases account for the top three causes of mortality, in the next 20 years COPD will become the 3rd leading cause of death in the United States [[1]]. Patients with advanced COPD are typically elderly and have significant co-morbid illness often also related to their tobacco smoking history. It is not surprising that they have high rates of depression but the underlying mechanisms leading to such depression are certainly multiple and may be a unique element of their respiratory illness. We will discuss the current estimates of the prevalence rates for depression in COPD, review the current understanding of the causes for depression in COPD patients, and present an approach to the assessment and treatment of depression in this patient population.

Prevalence of Depression in COPD Patients

Numerous studies have reported higher rates of depression and anxiety in COPD patients than those seen in the general population [2-7] however the quality of these studies has been highly variable. Van Ede [[4]] performed a systematic literature review up to 1997 and identified 68 studies for review of which only 10 met their criteria to warrant inclusion. Studies had to meet two of five criteria that included case-control design, random selection of patients in both groups, data to allow calculation of prevalence rate, exclusion of important physical co-morbidity, response rate of 80% or more. Of the 10 studies selected only 4 were case control studies. The remaining six studies were uncontrolled studies but accepted because they met at least two of the other four criteria.

While the two best-designed case control studies did show an increased prevalence of depression in COPD (7% vs. 1%) [[4]] and (28% vs. 19%) [[4]], the differences were not statistically significant. The case control studies that did show increased rates of depression had severe oxygen-dependent COPD [[4]]. The remaining uncontrolled studies from this systematic review reported prevalence rates between 6 and 42%, five of the six remaining studies had prevalence rates above 12% [[4]]. Nine of the 10 studies included patients with moderate to severe disease; the remaining study included patients with mild-to-moderate COPD.

The conclusions of this systematic review were that the empirical evidence for a significantly increased risk for COPD patients to develop depression is inconclusive and that the quality of the studies examining this important question is sub-optimal [[4]]. Subsequent epidemiologic studies have suggested that depression is present in 40% of COPD patients, compared to a rate of 15% in the general population [5-7]. The best of these studies utilize “examiner-driven” diagnostic tools to ensure that all research subjects meet established clinical criteria for a major depression, such as those of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) [[8]], rather than simply utilizing subject self-report of depressed mood.

In studies that have examined depressive disorders in diverse chronic illnesses, COPD patients suffer from depression with greater frequency and greater chronicity of the mood symptoms [9-15]. Because of the high level of overlap between many symptoms of depression and symptoms of COPD there has been concern that these reported rates of depression may be inflated. Yet, when direct comparisons were made between studies that utilized standard instruments versus those that used instruments that eliminated these confounding symptoms, the rates of depression in COPD were unchanged [[15]].

Relationship Between COPD and Depression

Depression not only increases the suffering of COPD patients but it also decreases function, impairs overall health status and rehabilitation efforts, and contributes to the ongoing use of tobacco [[7]][16&17].

The varied hypotheses as to why depression arises so frequently in COPD patients reflect an evolving understanding of the complex relationships between these two disease entities. Early theory held that depression was a “reaction” to the losses imposed by the illness [[18]]. Dudley described depressive symptoms as a mechanism by which the patient with COPD attempted to protect himself from environmental activity and exertion that might worsen respiratory symptoms [[19]]. Post attributed the development of depressive symptoms to patients' difficulty in managing their grief responses due to their declining function across multiple arenas of life [[17]]. Certainly COPD can impose staggering losses for COPD patients including loss of employment and occupational identity, disrupted family relationships, and declining independence and mobility. However, many of these same losses occur in other chronic illnesses, yet COPD patients demonstrate higher rates of depression. Accordingly, more recent theorists have begun to appreciate the complexity of the relationship between depression and COPD beyond psychological loss and grief.

One approach to this question has been to differentiate between early and late onset depression. “Early-onset” depression is defined as the presence of depressive symptoms prior to the diagnosis or symptomatic manifestation of COPD. How early depression presents obviously varies between individuals, but there are data to suggest that adolescents and young adults who are depressed are more likely to develop dependence on nicotine [20-22]. Starting smoking at an early age increases the duration of exposure and thereby the risk for later development of COPD. In adults, a history of past or recent depression has been associated with both increased difficulty with smoking cessation and an increased likelihood of recurrence of the depression if nicotine is discontinued [23-25].

In part, the association between depression and nicotine dependence may be due to components of cigarette smoke acting as an antidepressant, i.e., the act of smoking is a form of “self-medication” for an endogenous depression. There is evidence that nicotine inhibits monoamine oxidase A [[26]], an enzyme responsible for the intraneuronal breakdown of the monoamine neurotransmitters serotonin, norepinephrine, and dopamine. This inhibition of the enzyme is the same mechanism of action of some antidepressants. Cessation of smoking may put a smoker at risk for the re-emergence of a “nicotine-treated” depression.

“Late-onset depression,” in this context, is a depression that develops after a patient has been diagnosed with COPD. Here the traditionally proposed etiologies range from the incompletely processed grief as discussed before to the negative impact of certain COPD medications on mood, e.g., corticosteroids [[27]]. More recent research has begun to look at the direct effects of COPD on the brain and in so doing has added support to the concept that the derangement of cerebral microvascular physiology and chronic hypoxemia may both be involved in the increased susceptibility to co-morbid depression in these patients. Specifically, Brown and associates found that elderly patients, without respiratory illness, presenting with an acute diagnosis of depression, had greater MRI evidence for periventricular and subcortical white matter hyperintensities when compared to patients without psychiatric illness [[28]].

These same brain abnormalities appear to differentiate early- and late-onset depressed COPD patients, with the latter demonstrating a higher incidence of the finding [[29]]. Smoking and a reduced FEV₁ have been found to be independently predictive of the severity of this form of white matter disease [30&31].

The chronic hypoxemia seen in COPD patients also contributes to this vascular damage and has been found to disrupt the synthesis, release, and replenishment of a variety of neurotransmitters [[32]]. These specific pathological processes in COPD, as well as long-term exposure to cigarette smoke, while not entirely explanatory, are likely contributors to the development of depression, and may help explain the very high rates of depression found in patients with COPD. In examining the outcome of antidepressant therapy in such patients, it is crucial to consider that these differing etiologies, i.e., “early-” vs. “late-” onset of depression, may portend differences in treatment response.

Impact of Depression on COPD Patients

Multiple studies have demonstrated that when severity of respiratory symptoms is controlled for, COPD patients with co-morbid depression report lower quality of life and evidence greater objective impairment in functional performance when compared to non-depressed COPD patients [[7]][33&34]. In fact, in both the Felker [[33]] and Kim [[34]] studies, depression (as well as anxiety in the Kim study) accounted for more of the variance in the functional performance of subjects than did disease burden or COPD severity. Attempts to address these deficits with pulmonary rehabilitation demonstrated improved exercise tolerance, but no significant improvement in mood measures. Depression, by decreasing motivation and increasing sense of hopelessness, can decrease adherence to COPD treatment. For example, depressed patients are often less willing or able to participate in physical rehabilitation. With increasingly fragile self-esteem patients' commitment to using their oxygen may decline because of its impact on their appearance and their subjective sense of how they are seen. Appropriate management of depression in COPD will improve patient treatment adherence and outcome.

Management of Depression in COPD

Clear delineation of an effective treatment algorithm for co-morbid depression in COPD would improve the subjective experience of these patients, improve rehabilitation efforts in the management of COPD, and improve smoking cessation rates. Unfortunately, reliable and timely studies defining such an approach have not been completed. Most studies in the literature are either of poor quality, too short duration, or utilize older medications, limiting their current usefulness.

Of these studies, only Borson et al. [[35]] demonstrated that antidepressant medication was effective in reducing depression; all the other studies were inconclusive. Table 1 summarizes these studies.

Table 1.  Previous Studies of Antidepressants in COPD.

Author, Year	Agent, Dosing	# Subjects	Outcome measure	Duration	Results
Gordon (1985)	Desipramine 25–100 mg Ss @ 100 mg × 4 weeks	13, only 6 remained by end of study	Beck and Zung (Self-rated)	8 weeks	Improved depression scores with both placebo and desipramine
Light (1985)	Doxepin @ 100mg/d	12	Beck Depression Inventory (Self-rated)	6 weeks	Scores with no significant change from baseline
Borson (1991)	Nortriptyline, target of 1 mg/kg. Started at ¼ this dose then increased weekly to target	36	Hamilton Depression Rating Scale (Examiner driven) Also evaluated anxiety.	12 weeks	Nortriptyline use improved depression and anxiety sxs as well as subjective function
Smoller (1996)	Sertraline 25–100 mg/d	7 Case reports	Various	4 weeks to 7 months	Six out of the seven improved, raises question of subjective improvement in dyspnea
Evans (1997)	Fluoxetine 20 mg/d	76, randomized 38 to each arm; not exclusively COPD, multiple medical dxs	Hamilton Depression Rating Scale (Examiner driven and “computerized”)	8 weeks	No significant difference, but active group “trending toward improvement” at end of study period while placebo group scores remained static
Yohannes (1999)	Fluoxetine 20 mg/d	Started with 14, drop-out ended with 7	Montgomery Asberg Depression Scale (Examiner driven)	6 months	4/14 treated subjects no longer met criteria for Major Depressive Disorder

SXS = symptoms; DXS = diagnosis.

Unfortunately there were several recurrent methodological problems, including low power, insufficient treatment duration to allow for full antidepressant effect, excessive dependence on patient-report instruments as outcome measures for treatment effect on depression, low maximal doses of medications, and the utilization of older medications (e.g., tricyclic antidepressants) that are frequently poorly tolerated, if not overtly contraindicated in this population. Antidepressants routinely take 4–6 weeks to cause the cellular transformation associated with their antidepressant effect, so in a number of the studies inadequate treatment duration may be interpreted as lack of effect. In the more recent studies, Selective Serotonin Reuptake Inhibitors (SSRIs) were used and at generally effective doses.

In the best of the studies conducted to date, Borson treated 36 individuals with clinician-confirmed depressive diagnoses, 83% of whom also had significant anxiety symptoms [[35]]. Six patients did not complete the study, three dropped out due to medication side effects. At the time of this study over 10 years ago, Borson appropriately decided to use nortriptyline because it was the existing treatment of choice in medically ill elderly patients and because of the ability to verify dosing by checking blood levels. Baseline and outcome measures included state-of-the-art, examiner-driven assessments as well as subject self-report. Once targeted dosing levels werereached the patients were continued in treatment for 8 more weeks, allowing sufficient time for medication response. Borson reported that nortriptyline was significantly superior to placebo in improving mood and anxiety. Other findings included little impact of the antidepressant on dyspnea, either at rest or during structured exercise, or on physiologic function, as measured by spirometry and blood gases.

Nortriptyline-treated patients did rate themselves as globally less impaired on the Pulmonary Functional Status Instrument (PFSI) on 3 out of 5 dimensions, compared to 0/5 for the placebo subjects. Similarly there were significant improvements on the Sickness Impact Profile's (SIP) three major scales of overall functional disability, physical disability, and psychosocial disability for the active treatment group, none for the placebo group. The author concluded that “nortriptyline was an effective antidepressant and anxiolytic for moderately to severely disabled COPD patients” and that “tolerance for activity-related dyspnea was enhanced by treatment.”

An unpublished study from Eiser and colleagues from University Hospital Lewisham in London apparently demonstrated significant improvements in self report and psychiatricevaluation of depression after a 6-week course of the SSRI paroxetine [[36]].

Based on this review, we conclude that there has been inadequate study of the efficacy of treatment for depression in patients with COPD. The most comprehensive and best-designed study, Dr. Borson's, is now out of date because tricyclic antidepressants are no longer the first-line medications used in depressed patients. Most of the studies that have examined newer SSRIs, such as sertraline, have “fatal” flaws. The one study that has examined an SSRI appears to support the use of paroxetine, but these data need to be subjected to the scrutiny of full, peer-reviewed publication, then replicated. Furthermore, the applicability of other SSRIs also needs to be evaluated.

Recommendations

In the face of limited guidance from the research literature the question remains as to how to design the best management of depression in COPD patients. The standard precepts of all good medical care—primum non nocere, careful diagnosis, consideration of contributing factors, thoughtful selection of pharmaceuticals, and utilization of consultation when appropriate—still hold. Table 3 summarizes the management of depression in COPD patients.

The first step in managing depression is confirmation of the diagnosis. This requires a thorough history of the patient's mood symptoms as well as a review of the collateral symptoms of depression including level of self-esteem, ability to concentrate, subjective sense of guilt, and variations in appetite and sleep patterns. As discussed previously, careful discrimination between those symptoms particularly indicative of depression, i.e., poor self-esteem, a sustained inability to brighten, and of course, overt suicidal ideation, and those also common with COPD—fatigue, decreased motivation, disrupted sleep and appetite—is an important first step. Next, completing select labs such as thyroid function tests, CBC, and vitamin B12 and folate levels allows one to rule out some of the most common somatic etiologies for symptoms of depression.

Once the diagnosis of depression is made, it is useful to clarify if the patient is more likely suffering from an “early-” or “late-onset” depression to guide adjunctive treatments. In addition to asking when the patient first began experiencing depressive symptoms there are other clinical clues as to which phenomenon the patient might be struggling with. To the extent that late-onset depression is correlated with greater global CNS dysfunction (associated with increased white matter hyperintensities) we might expect that these patients will manifest more associated cognitive dysfunction, functional disability, limited insight and psychomotor retardation, and are more reliably non-psychotic. In turn, patients with early-onset depression are likely to have a more “classical” presentation of depression, but in the context of COPD one should expect this group to have greater difficulty with smoking cessation. For those patients with primarily a “late-onset” picture increasing support networks and taking steps to protect against vascular damage will be especially important, while “early-onset” patients may require more structure and support in smoking cessation programs.

In addition to the supportive elements of the treatment plan, a physician working with a depressed COPD patient will also want to consider the prescription of an antidepressant. While this seems a fairly straightforward process it must be remembered that this population is more susceptible to depression than other groups, and this implies a more complicated disease state. Among the other factors previously discussed (disease impact on CNS, genetic susceptibilities to both smoking and depression) it is also true that this population suffers from a higher frequency of anxiety and panic disorder than others [[37]]. This susceptibility to panic and anxiety—along with frequent medical co-morbidities, increased risk for medication interactions and general patient debilitation—contributes to making management of depression in this group much more challenging than in a general psychiatric clinic patient, the population for whom the protocol algorithms were designed. So, while the standard algorithms can be used for guidance, it is imperative to recognize their limitations in applicability to treatment planning, including antidepressant selection, for this population.

The available research does not clearly support the use of any particular medication over others. SSRIs are the first-line treatment for primary depression and it is reasonable to start with them in this scenario as well. For now choosing between them is a function of side-effect profiles, co-morbid diagnoses that might benefit from secondary indications, medication interactions and the impact of half-life. Table 2 provides a comparison of the SSRIs that should help with the selection of an antidepressant. The side effects potentially common to all SSRIs include difficulties during initiation of therapy with GI upset, diarrhea, headaches, tremors, and either psychomotor activation or sedation. This latter continuum, i.e., from sedation to agitation, is frequently problematic for COPD patients in that their dyspnea tends to impose a baseline level of anxiety that can be aggravated by SSRIs, leading to impaired adherence. Over time the SSRIs also carry a risk for sexual dysfunction, in particular diminished libido and delayed or inhibited climax. Side effects that are more unique to individual agents are shown in Table 2.

Table 2.  Comparison of Selective Serotonin Reuptake Inhibitors.

Medication	Approved indications	Most common side effects	Half-life	Usual daily dosing
Fluvoxamine (Luvox)	OCD	Significant GI upset, requires BID/TID dosing	15 hours	50–150 mg PO BID
Fluoxetine (Prozac)	Depression, OCD, Bulemia	Minimum weight gain	2–3 days met 7–9 days	20–40 mg PO daily; higher doses in eating disorders (80 mg)
Sertraline (Zoloft)	Depression, OCD, PMDD, PTSD, Panic		26 hours, no mets	50–200 mg PO daily
Paroxetine (Paxil)	Depression, OCD, Panic	High risk for sexual side effects, intense discontinuation syndrome, + weight gain, most often sedating	21 hours, no mets	20–50 mg PO daily
Citalopram (Lexapro, Celexa)	Depression, GAD	Minimal weight gain; occasional tremors		Celexa: 20–60 mg PO daily; Lexapro: 10–40 mg PO daily
Venlafaxine (Effexor) Also available in XR formulation.	Depression, GAD	Frequent agitation at start-up, risk of hypertension particularly with doses > 300 mg daily		XR: 75–300 mg PO daily

A full discussion of all possible medication interactions involving SSRIs and the range of medications that are used for COPD patients is beyond the scope of this paper, but we can draw some general outlines. The SSRIs as a group are substrates of P450 2D6 primarily and 3A4 and 2C19 secondarily. In addition they are also mild inhibitors of 2D6 and 2C19. As such the SSRIs are typically not very problematic in interaction with the majority of bronchodilators or steroids. An exception to this broad sweep is theophylline, which is notorious for its tendency to interact with other medications, in this case particularly with fluvoxamine. The pharmacokinetic interaction between fluvoxamine and theophylline can decrease the clearance of theophylline by up to 80%, increasing the steady state serum concentrations of the bronchodilator 2 to 3 times [[38]].

Antibiotics also present a potential risk for interaction with the SSRIs, particularly the macrolides and the fluoroquinolones. These two groups of antibiotics both block potassium channels leading to possible QTc prolongation [[39]]. SSRIs have also been found to have potential for the same cardiotoxicity [40&41]. As is the usual case with this phenomenon, the risk associated with a single medication may be marginal, but prescribers must be attentive to amplifying this risk by combining offending agents. Finally, the macrolides tend to inhibit the P450 3A4 enzymes with varying degrees of potency. Since the SSRIs are substrates of this enzyme chain, one should be aware, theoretically at least, that this potential interaction might be a source of increasing antidepressant concentration, although this has not been reported in the literature.

Table 3.  Algorithm Approach to Depression in COPD Patients.

Whichever antidepressant is chosen, standard protocol requires that it be given a full trial before it is concluded that the medication is ineffective. A full trial implies that the dosing is titrated up (as long as full remission of symptoms hasn't been acheived) to the top end of usual daily dosing and that each dosing level is maintained for a sufficient duration to allow the cellular changes associated with antidepressant effect, usually 4 to 6 weeks. In COPD patients their increased susceptibility to anxiety often requires a slower titration and more frequent monitoring over the early weeks of a trial.

Standard psychiatric algorithms recommend that if the first SSRI is ineffective a second should be tried. If, after a complete trial, the second antidepressant also proves ineffective, then trials of medications from the “second-line” should be considered. These include venlafaxine, mirtazapine, bupropion, and the tricyclic antidepressants. Each of these brings particular challenges to the COPD patients. Venlafaxine, while approved by the FDA for both depression and generalized anxiety, is highly anxiogenic in the start-up phase, so much so that it is often very difficult for COPD patients to initiate. Bupropion, while less likely to cause sexual side effects, also often increases agitation and impairs sleep. Mirtazapine also typically causes fewer sexual side effects, but is associated with significant weight gain. As noted above, the tricyclics do have some demonstrated benefit in this population but also confer significant anticholinergic and hypotensive side effects as well as presenting a significant risk for lethality in overdose. Beyond monotherapy there are combination therapies that combine antidepressants and combinations that utilize adjunctive medications. In these scenarios synergistic gains must be balanced against increased risks for side effects, medication interactions and, frequently, patients' dismay at “yet another medication.”

In addition to medication trials, patients may also benefit from the addition of appropriate psychotherapy. Cognitive behavioral therapy allows for a very problem-focused approach to managing the daily challenges these patients face. Other options include supportive or psychodynamic therapy depending on the needs and interest of the patient. It is also critically important to recognize the impact both COPD and depression have on family and friends around a patient and to include “support for the support group.”

Finally, for patients whose depression is refractory to all other interventions electroconvulsive therapy can be life saving.

When to Refer to Psychiatry

The complexity of effectively managing depression in COPD patients raises the question of when a patient should be referred to a psychiatrist. Some clear indications that such a consultation would be useful include a patient with multiple co-morbid psychiatric diagnoses, a patient with manic-depressive illness or other types of bipolar depression, a patient with depression that has shown itself to be largely refractory to treatment options with which the primary physician is comfortable. Of course, the most appropriate answer to the question of when to request a consult is whenever the treating physician feels it would be useful. Because depression is so common in COPD patients a psychosocial evaluation should be routinely included at initial assessment and ongoing maintenance visits should include monitoring of mood. Because the risks for refractory illness, side effect complications and patient difficulties can be so great in COPD patients, pulmonary physicians and primary providers who care for COPD patients should have a lower threshold for involving their psychiatric colleagues.

Future Research Needs

The question of how best to treat our patients struggling with both COPD and depression is critically important and requires more rigorous scientific research than has occurred to date. That research should address not only the question of “which antidepressant” but also explore why these two illnesses so often travel together, how the contributing etiologies impact treatment planning and the selection of the best concurrent treatment psychotherapies. Another area of developing research is further elucidation of the CNS structural changes discussed above on MRI and their relation to affective disorders. Are they merely associated or do they point in the direction of causality? In COPD patients are there other clinical differences between early- and late-onset depression and do these two separate entities also warrant different psychopharmacologic treatment? In addition to the high co-morbidity of COPD and depression, panic and anxiety are also frequently seen with COPD, both in the presence of and absence of a diagnosis of depression. Again understanding the relevance of these co-morbidities and how they impact disease management needs to be explored. Thoughtful research is necessary to improve both their quality and quantity of life.

Acknowledgments

The author would like to acknowledge Glen Whelan, Pharm. D., for his above-and-beyond level of help with the pharmaceutical tangles of the P450 system.