Prevalence and outcome of partial remissionin depression

Abstract

The goal of treatment of major depression should be full remission. Many patients, however, fail to achieve or maintain symptom-free status. Residual depressive symptoms are common, even where there has been a robust response to antidepressant therapy. In clinical studies, approximately one-third of patients achieve a full remission, one-third experience a response and one-third are nonresponders. Partial remission is characterized by the presence of poorly defined residual symptoms. These symptoms typically include depressed mood, psychic anxiety, sleep disturbance, fatigue and diminished interest or pleasure. It is currently unclear which factors predict partial remission. However, it is clear that residual symptoms are powerful predictors of relapse, with relapse rates 3–6 times higher in patients with residual symptoms than in those who experience full remission. Residual symptoms are also associated with more medical and psychiatric visits, increased public assistance, disability benefits, thoughts of and attempts at suicide and chronicity. The risk of stroke and coronary events is also higher in patients with residual depressive symptoms. The substantial proportion of patients who achieve only partial remission has traditionally been neglected in antidepressant trials. Given that residual symptoms may relate, in part, to an incompatibility between patient and treatment, further research is needed to predict a better match. These symptoms are a clinically relevant state of illness, and the correct choice of initial antidepressant treatment should offer the greatest chance of achieving full remission.

Introduction

Despite many effective therapeutic strategies, depression continues to be a highly prevalent, disabling and costly condition.^1,2,3 The World Health Organization identifies unipolar major depression as the leading cause of disability in the world.⁴ Although most patients experiencing an episode of depression undergo some acute improvement after treatment, long-term outcomes remain disappointing.^5,6 Although full remission is the goal of treatment,⁷ many patients fail to achieve or maintain symptom-free states.⁸ The exact nature and origin of residual symptoms is currently debated, but their tremendous impact on outcomes such as future relapse, morbidity and mortality is clear. A better understanding of residual symptoms may inform future treatment choices in depression and ultimately improve prognosis.

For this review, a MEDLINE search was conducted using the terms “residual symptoms of depression,” “partial remission from depression” and “subsyndromal depression.” Further key references were then identified from the initial review papers. As well as focusing on current theories regarding residual symptoms, this review will also highlight inadequacies in the existing literature.

Partial remission

The optimal outcome in the treatment of major depression should be the virtual elimination of symptoms and a return to a premorbid level of functioning. However, clinical depression with or without treatment can result in various negative outcomes including chronicity, relapse and recurrence.^5,9 Definitions of response, remission, relapse and recurrence are consistent with those described in the introductory article to this series (i.e., those of Frank et al¹⁰). Another important adverse outcome for major depression is partial remission with residual symptoms. Frank and colleagues¹⁰ defined partial remission as a period of sufficient improvement such that an individual no longer fulfills the criteria for major depressive disorder (MDD) but continues to evidence more than minimal symptoms. Partial remission with minimal or residual symptoms has been defined in numerous ways in the literature, most commonly using the Hamilton Rating Scale for Depression (HAM-D) (Table 1).^{5,11,12,13,14,15} Other tools have also been used to evaluate residual symptoms such as the Clinical Interview for Depression (CID) and the Montgomery–Asberg Depression Rating Scale (MADRS), but to facilitate comparisons, this review will focus on studies using the HAM-D.

Table 1

Subsyndromal or subthreshold depression, a related phenomenon (less than threshold symptoms for MDD), is a much broader category as it includes spontaneous depressive symptoms in community studies, prodromal symptoms and residual symptoms.^16,17 Although the literature suggests similarities to partial remission (e.g., negative impact, increased episodes of depression), this concept is outside the scope of this review.

The existence of partial remission has been widely recognized, but its significance in clinical practice has not.¹⁸ Clinical trials report rates of response (50% reduction in symptoms), nonresponse and, in a minority of studies, remission (HAM-D ≤ 7).¹⁰ Partial remission and residual symptoms are not reported in most trials designed to look at the efficacy of antidepressant strategies. In addition, most drug trials last 6–8 weeks, not long enough for most patients to achieve remission.

In most studies specifically designed to determine the nature and prevalence of partial remission, patients are considered to have residual symptoms if they responded to therapy but had a HAM-D score of 8 or more (Table 1).⁵ However, it has now been suggested that patients who achieve full remission, as defined by even the most conservative criteria, may continue to have residual symptoms.^10,13 Estimating the impact of residual symptoms is further limited by the fact that there are few known baseline HAM-D scores for the general population. Therefore, accurate comparisons are not possible.

Prevalence of residual symptoms

Many studies have reported on the prevalence of symptomatic patients after various pharmacotherapies and psychotherapies, both prospectively and naturalistically, in psychiatric clinics and in primary care.^{10,11,13,19,20,21,22} However, each study has a unique method for reporting residual symptoms, and the patient populations, treatment modalities and time periods differ. Although trials of antidepressant treatment typically report response rates in the order of 70%–80%, longer-term follow-up studies have shown that no more than a third of patients achieve full remission from their depressive symptoms.²⁰ The report by Kupfer and Spiker²³ is representative of most studies looking at long-term response to treatments for depression, suggesting the “rule of thirds” may apply: a third of patients achieve complete remission, a third do not respond and a third show partial remission.²³

In a long-term outcome study over 4 years, 50% of patients who had been treated with amitriptyline and psychotherapy experienced residual symptoms.¹⁹ Similarly, in a 4-year follow-up study of elderly depressed patients, 38% experienced residual symptoms 1 year after remission and 20% had residual symptoms after 4 years.²¹ Paykel and colleagues,⁵ who followed a cohort of 64 inpatients and outpatients who had achieved partial remission during treatment for major depression, reported that after 15 months, 32% of patients exhibited residual symptoms. Finally, in a 10-year follow-up, the National Institute of Mental Health (NIMH) Collaborative Depression Study Group reported that 34% of patients were in partial remission.²²

Patients who meet remission criteria may still experience residual symptoms. In a study of subjects who were in full remission (i.e., HAM-D ≤ 7) after treatment with fluoxetine (20 mg for 8 weeks), more than 80% of patients had 1 or more symptoms, and more than 30% had more than 3 residual symptoms of MDD.¹³

Nature of residual symptoms

In a study of patients who remained partial responders to antidepressant therapy (HAM-D score 8–18), 47% reported depressed mood, impairment of work and activities, psychic anxiety or sexual dysfunction, to at least a moderate degree.⁵ Symptoms that are more commonly associated with severe depression, including late insomnia, retardation, agitation, hypochondriasis, weight loss and loss of insight were less common. The most persistent residual symptoms reported in a study of elderly patients (mean age 67 yr) during the continuation phase of treatment were apathy, anxiety (both psychological and somatic), anergia, insomnia, feelings of guilt and loss of libido.¹⁴ Although sexual dysfunction was reported in some studies surveyed, drug side effects were not systematically excluded from residual symptom scores.^5,14

The 3 most common residual symptoms reported in a study of patients in full remission were sleep disturbances (44%), fatigue (38%) and diminished interest or pleasure (27%).¹³ Depressed mood and suicidal ideation were rarely reported. In another study of patients in full remission, the most common residual symptoms were generalized and somatic anxiety and irritability.²⁴ When patients in remission were compared with a group of volunteers who had never been depressed, they demonstrated significantly more problems with social function, problem solving and dysfunctional attitudes.²⁵

Various models have been proposed in an attempt to explain the cause of residual symptoms. A “vulnerability” model suggests that certain pre-existing personality traits are a risk factor for depression and that these traits persist after recovery.²⁶ In contrast, the “scar” model proposes that depressive episodes cause lasting changes in personality.²⁷ Several investigators have reported that neuroticism-like personality factors predispose to the development of major depression,²⁸ whereas extroversion-like factors have been associated with a better response to therapy.^26,29 This suggests that in some patients, the presence of “residual symptoms” may represent a return to baseline personality characteristics, which are also those that predispose to depressive illness. Alternatively, residual symptoms may represent persistent illness; that is, the original illness continuing in a milder form.⁵

Predictors of residual symptoms

Attempts to define predictors of partial and full response to antidepressant therapies have yielded conflicting results in terms of baseline severity,^5,13,30 personality traits^5,14 and the impact of life events.^5,13,14 Paykel's group investigated a number of patient characteristics and found that only severity of illness was a predictor of residual symptoms.^5,30 Conversely, for patients in “full remission” (HAM-D₁₇ ≤ 6), Nierenberg and colleagues¹³ reported that the presence of residual symptoms was not predicted by baseline severity of depression. Some studies have found no relation between residual symptoms and life events.^5,13 However, Opdyke et al¹⁴ reported that residual symptoms during continuation treatment in elderly patients who were in full remission (i.e., HAM-D₁₇ 10 or less) were observed more often in subjects with associated severe life events or ongoing major stressors.¹⁴ Similarly, personality traits have been associated with residual symptoms in some studies but not in others.^5,14

No relations have been found between residual symptoms and sociodemographic factors, family and personal history, follow-up care, comorbid conditions, chronic medical burden, social support or past or present illnesses.^5,13,14 In addition, residual symptoms are prevalent, not only in patients who receive psychotherapy, but also in those treated with pharmacotherapy.¹²

Overall, residual symptoms cannot be accurately predicted by age, sex, marital status, number of prior episodes, duration of the current episode, treatment courses or comorbid conditions, but the severity of depression, life stressors and personality dimensions have some predictive value.

Residual symptoms and relapse

Relapse and recurrence are important and, unfortunately, frequent long-term outcomes in the management of patients with depression. The presence of residual symptoms has been associated with a significant increased risk of relapse after treatment with either pharmacotherapy or psychotherapy.^{5,12,31,32,33}

Thase and colleagues¹² found that relapse occurred in 52% of the patients who had a partial response to cognitive therapy (i.e., HAM-D 6–10 for 2 consecutive weeks) but occurred in only 9% of those who had a full response to treatment (i.e., HAM-D ≤ 6 for 8 consecutive weeks).¹² In Paykel's study of those who had responded to treatment (i.e., HAM-D 8–18), 76% of those with residual symptoms, as opposed to 25% of those without, relapsed over the 12- to 15-month follow-up period.⁵

In the NIMH Collaborative Depression Study, patients with (n = 82) or without (n = 155) residual depressive symptoms after treatment were followed naturalistically for 10 years or longer.²² Patients who had residual symptoms relapsed 3 times faster than patients who achieved full remission. A history of recurrent episodes has also been associated with higher relapse rates, but, in this study, the risk for relapse associated with residual symptoms was significantly greater (odds ratio 3.65) than that associated with a history of recurrent MDD (odds ratio 1.64). Partial remission was still highly predictive of relapse in this study, despite the more stringent criteria for full remission than those proposed by Frank and colleagues.¹⁰

In a 1-year follow-up of patients who had received treatment for the index episode of unipolar depression, 50% relapsed;³¹ those who relapsed showed higher levels of residual symptoms and Clinical Global Impressions Scale scores that described greater impairment. Neither the HAM-D score at index episode nor at recovery were predictive of relapse. In another study, patients who received psychotherapy (n = 20) to reduce residual symptoms had lower rates of relapse than those who did not (n = 20) after 2 years (15% v. 35%, p ≥ 0.27), 4 years (35% v. 75%, p < 0.05) and 6 years (50% v. 75%, p = 0.06).^24,32,34 However, the differences were significant at 4 years only.³⁴

A subset analysis compared relapse rates of patients with more severe residual symptoms (HAM-D > 12) with those with milder symptoms (HAM-D 8–12). Surprisingly, the rate of relapse was higher among patients with milder symptoms (90%) compared with those with severe symptoms (57%).⁵ It is unknown whether relapse is related to specific symptoms or to the number of symptoms.

The importance of partial remission as a prognostic factor after treatment of depression appears to be independent of the treatment modality adopted. Residual symptoms appear to predict relapse after treatment with pharmacotherapy³⁵ or psychotherapy.^12,36

The costs of residual and untreated symptoms

In addition to a higher risk of relapse, residual symptoms are associated with a number of other negative outcomes. Residual symptoms after recovery from depression are associated with more medical and psychiatric visits, emergency room use, psychiatric hospital admissions, increased public assistance, disability benefits, thoughts of suicide and attempted suicide.^2,22 Chronicity is also increased in patients with residual symptoms. A 12-year follow-up of patients after their first major depressive episode demonstrated that those with residual symptoms during recovery had more severe and chronic future courses.³⁷

Untreated depressive symptoms also predict future cardiovascular disease and stroke. The Montreal heart studies demonstrated that the risk of cardiovascular mortality after myocardial infarction was significantly increased in patients with a score 10 or more on the Beck Depression Inventory (OR was 3.0 at 12 mo and 7.82 at 18 mo).^38,39 Depression was identified in over 40% of patients with unstable angina and was associated with a significantly increased risk of cardiac death or nonfatal myocardial infarction at 1 year, even after controlling for other significant prognostic factors (adjusted OR 6.73).⁴⁰ The Stockholm Female Coronary Risk Study included 292 women, aged 30–65 years, who had been admitted for an acute coronary event between 1991 and 1994.⁴¹ After 5 years of follow-up, 35% of the women who lacked social integration and had 2 or more depressive symptoms had experienced a relapse of their coronary disease, recurrent acute myocardial infarction or cardiovascular death compared with only 9% of the women who were free of depressive symptoms and had good social integration. In a population-based cohort of 6095 adults, depressive symptoms were also found to be predictive of stroke.⁴² After adjustment for other risk factors, the presence of self-reported depressive symptoms was associated with a relative risk of stroke of 1.73.

Management of residual symptoms

Residual symptoms are a tremendous economic burden to the health care system and represent a clinically relevant state of illness in people who suffer from MDD. Therefore, it is important to identify treatment strategies that minimize the incidence of residual depressive symptoms. The drug selected for initial treatment should offer the best chance to induce a full remission. Data suggest that amitriptyline and venlafaxine may offer higher rates of full remission.^42,43,44,45 Unfortunately, a relatively high incidence of adverse events limits the routine use of the tricyclic antidepressants. It should also be noted that the meta-analysis of venlafaxine by Thase and colleagues⁴⁵ was company sponsored.

Cognitive therapy (CT) has been shown to benefit patients who have only a partial remission with antidepressant treatment.⁴⁶ Patients randomized to continue pharmacotherapy with the addition of CT had significantly reduced relapse rates compared with those who continued pharmacotherapy alone (29% v. 47%). In a long-term prospective study, patients receiving continuation therapy with a modified version of CT designed to address residual symptoms after antidepressant treatment had a significantly lower level of residual symptoms and a lower rate of relapse over 6 years of follow-up.^24,32,34 Similar results were reported in patients with recurrent depression, where those receiving continuation therapy with CT had a significantly lower level of residual symptoms, and at the 2-year follow-up had a much lower rate of relapse (25%) than the clinical management group (80%).⁴⁷

Results from acute combination treatment with both pharmacotherapy and psychotherapy for depression have been inconsistent.⁴⁸ However, a study by Keller and colleagues⁴⁹ suggests that, in addition to increasing overall response rates, combination treatment with nefazodone and the cognitive behavioral analysis system of psychotherapy significantly increases rates of complete remission (HAM-D < 8) compared with drug therapy or psychotherapy alone.⁴⁹

Summary

Despite the inconsistencies in the research concerning partial remission — the definition being first and foremost — it is clear that residual depressive symptoms are associated with a range of clinically significant negative outcomes. Even patients with minimal levels of residual symptoms are at greater risk of early relapse into depression. They are burdened with greater levels of social dysfunction and experience higher rates of physical morbidity and mortality than patients who have achieved full remission. Residual symptoms may relate, in part, to an incompatibility between the treatment and the patient (personality factors) — further research is needed to allow a better prediction of the best match. It is clear that treatment strategies should be directed toward reducing residual symptoms.