Time course of response and remission during antidepressant treatment

b y H . – J . M ö l l e r, F. H . S e e m ü l l e r,
a n d M . R i e d e l , G e r m a n y

Hans-Jürgen MÖLLER, MD, Prof
Michael RIEDEL, MD, Prof

Department of Psychiatry
Ludwig-Maximilians University

Major depression is a complex and inhomogeneous illness with an etiopathogenesis that is based upon multiple factors that may act at different levels (psychological, biological, genetic, and social). It is a matter of fact that one single treatment, or one single treatment approach, can therefore never be adequate for such a diverse illness. There is hope that in the near future we will be able to treat patients on an even more individual level by using genetic or brain morphologic markers. In the case of drug resistance, pseudoresistance—with special attention given to adherence—should be ruled out first. Response to treatment can occur very early, even within the first 2 weeks of the treatment course. In clinical trials, the highest response rates appear within the first 4 weeks. When, exactly, and to what extent response appears, also depends on its exact definition. Two definitions have achieved acceptance in the research community—that of early improvement (20% reduction on the Hamilton Rating Scale for Depression [HAM-D] within the first 2 weeks) and that of response (_50% reduction of the initial HAM-D score). The first may be especially useful in monitoring early symptom alleviation due to the respective treatment, while the second may be more useful in detecting sustained changes, especially when combined with a time criterion. Both are highly predictive of remission, which is usually observed at a lower rate compared with response, but which in clinical trials surpasses the response rate from week 4 to 6 onward. Thus, response to treatment may be considered as the precursor of remission, ie, the phase inaugurating the natural healing process induced by antidepressant therapy and which, in the majority of cases achieves its goal, ie, remission, the complete relief of all symptoms.
Medicographia. 2009;31:118-125. (see French abstract on page 125)

Keywords: major depression; response; remission; time course; etiopathogenesis

Natural evolution
of a depressive episode

Mechanisms involved in the emergence of depressive episodes are multifactorial and not yet fully understood. As recently pointed out by Kenneth Kendler, a comprehensive etiological understanding of psychiatric disorders will require the integration of multiple explanatory perspectives.1 If we want to shed light on the complex interactions between the pathophysiology of depression, its treatment, and its cure, we need to consider all the different psychopathological subtypes, the contribution of different neurobiological and psychosocial etiopathogenetic factors, the clinical response to treatment, and the long-term prognosis, and additionally the comorbidity with other psychiatric or somatic disorders. There is enormous variation between individual patients with regard to these different factors, which corresponds to the variability of outcome regarding the time course of response and remission. In the following paragraphs, we will first try to summarize some of the most recent findings concerning the impact on de- pression of four major influential levels; namely, the psychological, environmental, biological, and genetic levels. An excellent overview of the most recent concepts in our understanding of the biological mechanisms of major depression can be found in reference2.

Stressful life events and gene-environment interaction
Stressful life events that occur during an individual’s lifespan influence the onset and course of major depression. In a recent paper by Kendler and colleagues,3 the relationship between dependent stressful life events (life events that result from the respondents’ own behavior) or independent stressful life events (events that mostly result from “bad luck”) and the onset of major depression was systematically investigated in a sample of 1898 female twins (monozygotic and dizygotic).3 Independent and dependent stressful life events were associated to a highly significant extent with new onset of depressive episodes, and the strongest association was observed for the dependent events. The odds for the onset of major depression in the month of a life event were 5.64 for all subjects.3 However, major depression can also arise without any prior life event and the vast majority of people experiencing life events do not develop depression. One reason for this might be that the impact of life on the development of an illness may be moderated by a geneenvironment interaction. In 2003, Avshalom Caspi prospectively investigated why life events lead to depression in some people but not in others.4 In a representative birth cohort of 1037 children who had had comprehensive psychiatric assessments at the ages of 3, 5, 7, 9, 11, 13, 15, 18, and 21 years, a functional polymorphism of the promoter region of the serotonin transporter gene (5-HTT) was found to moderate the influence of stressful life events on depression.4 Specifically, the study revealed three major results. First, with regard to the gene-environment interaction, it was found that life events after the 21st birthday significantly predicted a new onset of depression at age 26 among carriers of an S-allele who had no history of prior depression, but did not predict onset of depression among l/l homozygotes. Second, even suicide ideation, which is always susceptible to having an underlying biological mechanism, could be significantly predicted for individuals with the S-allele but not for l/l homozygotes (Figure 1). Third, childhood maltreatment during the first life decade also significantly predicted adult depression among S-allele carriers but not among l/l homozygotes. In addition, the sole relationship between genotype and depression not associated with a life event was not significant.4 In summary, this model further supports the concept that genetic variants with a high prevalence in the general population probably act to “promote an organism’s resistance to environmental pathogens.”4

Monoamine deficiency hypothesis
On a transmitter level, the monoamine deficiency hypothesis remains of extreme importance. Apart from the fact that all antidepressant agents have an effect on at least one of the monoamine transmitters (dopamine, serotonin, norepinephrine), there is also compelling evidence from the opposite direction. Tryptophan is rate limiting for serotonin synthesis in the brain. Oral tryptophan depletion does not induce depression in mentally healthy subjects, but will cause depression in patients successfully treated with a selective serotonin reuptake inhibitor (SSRI). Likewise, á-methyl paratyrosine does not induce depression in normal people, but does in patients who have been treated with a norepinephrine reuptake inhibitor.5 The importance of dopamine in depression is highlighted by the high comorbidity rates of depression in patients with Parkinson’s disease, the antidepressant effect of buproprione, and the depressogenic effect of dopamine depletion with reserpine.

Figure 1
Figure 1. Relationship between suicidal ideation and attempts, number
of stressful life events, and 5-HTT allele carrier status.
After reference 4: Caspi A, Sugden K, Moffitt TE, et al. Influence of life stress on depression:
moderation by a polymorphism in the 5-HTT gene. Science. 2003;301:386-389.
Copyright © 2003, The American Association for the Advancement of Science.

Hypothalamic-pituitary-adrenal axis
Another possible link between life stress and depression might be the moderation of depression through the hypothalamic-pituitary-adrenal axis. Depressed subjects can show elevated plasma cortisol levels6 and elevated corticotropin releasing hormone (CRH) levels in the cerebrospinal fluid.7 Furthermore, the normal cortisol suppression response is absent in about half of the most severely depressed subjects.8 Consequently, antidepressantinduced remission is also associated with the reversal of the abnormal suppression test, supporting the notion of a causal relationship.8 Most recent research for new antidepressant agents has also focused on glucocorticoid receptor blockers, with promising preliminary results.9 Higher glucocorticoid levels may also be associated with reduced neurogenesis, and have been linked to a decreased hippocampal size in patients with depression.10 Brain derived neurotrophic factor (BDNF), which plays an important role in neurogenesis involving neuronal plasticity, synaptic growth, and neuronal cell survival, was found to be significantly reduced in a cohort of suicide completers.11 Antidepressant treatment options such as regular physical activity,12 electroconvulsive therapy (ECT)13 or antidepressant compounds14 have been shown to significantly elevate BDNF levels.

Circadian rhythms
Circadian disturbances occur in depression and include not only disturbances to physiological parameters (eg, body temperature) or biological parameters such as cortisol secretion, but also mood and the sleep-wake cycle. Therapeutic treatments aimed at the regulation of these disturbances in depressed patients have been shown to have a significant impact in relieving symptoms of depression.15
It is not surprising given this diversity and complexity in the etiopathogenesis of depression and the pharmacological mechanisms of antidepressants that there is a broad range of possible answers to the question of which therapeutic approach is the most beneficial for the individual patient. For such a highly inhomogeneous and diverse medical condition, response and remission can hardly be achieved with only a single therapeutic strategy. Today, we still need to make use of the full range, variety, and combination of different therapeutic strategies, such as inpatient and outpatient treatment, the differential and sequential use of psychopharmacological compounds, pharmacological combination and augmentation strategies, as well as other highly effective biological treatments including light therapy, ECT and transcranial magnetic stimulation, and psychotherapy, in order to deal with this illness according to the rules of good clinical practice.16 Apart from subtyping depression and using predictors, which will be discussed later, good allies in that challenge remain time and patience to obtain the optimal treatment regimen for each individual patient. This notion finds support in naturalistic inpatient data from a large German inpatient study, as well as in the outpatient Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study.17,18
In both trials, the patients were actively involved in the decision-making process of their next treatment step, and results of both trials indicate that the longer the treatment and thus the time to individualize and optimize the treatment regimen, the better are the response and remission rates.17,18

Evolvement of a major depressive episode:
the psychopathological perspective

The natural evolution and time course over which depressive symptoms usually first develop has rarely been systematically investigated. However, better knowledge about early warning signs and the development of major depressive episodes could significantly improve health education and might lead to an improvement in early detection.
One of the most consistent early warning signs and prodromal symptoms of major depression is sleep disturbance.19-21 Sleep serves as an important timer for the daily circadian rhythm and is found to be disturbed about 5 weeks prior to the new onset of a depressive episode in patients with recurrent depression.20 Data from the National Institute of Mental Health catchment area program across four sites in the US found strong associations between sleep disturbances and the depressive symptoms of diminished sexual drive, feelings of worthlessness or excessive guilt, and trouble concentrating or thinking, 1 year prior to the new onset of a major depressive episode.21,22 In very fast developing episodes, loss of libido and suicidality can be the first symptoms to develop. This is of special importance, because loss of libido often tends to be overlooked clinically, since clinicians are often reluctant to ask questions about a patient’s sexual life. In addition, these results do suggest that suicidality should be given notably more attention in patients with a fast onset of a depressive episode.23 The core and illness-defining symptoms that develop in the further course of depression are depressed mood and/or the inability to experience joy (anhedonia).
Different depressive syndromes can develop, depending on which symptoms evolve. From a psychopathological perspective, the most important ones that we differentiate today, because of their different prognostic and therapeutic implications, are melancholic, atypical, and psychotic depression. Recently, Gin Mahli and Gordon Parker developed a model for the relationship between psychotic melancholia, nonpsychotic melancholia, and nonmelancholic depression, and the majormonoamines thought to be involved in depression. The authors postulate that in nonmelancholic depression with its primary affective symptoms, serotonergic dysfunction may prevail, while nonpsychotic melancholia, which is accompanied more by psychomotor disturbance, derives from a more pronounced norepinephrine dysfunction. With respect to psychotic depression, dopamine may be involved to a higher degree (Figure 2).24 As another example, the subtype of atypical depression can be mentioned here, which has long been seen as an indication for treatment with monoamine oxidase inhibitors.25,26
Finally, in a full-blown depressive episode, depressive symptoms appear in the following percentages (data from 1014 depressed inpatients using the 21-item Hamilton Rating Scale for Depression [HAM-D21]; given in declining order of percentage): depressed mood 99.4%, loss of interest in work and activities 99.2%, somatic symptoms 85%, psychic anxiety 84.3%, feelings of guilt 78.5%, insomnia middle of night 78.3%, genital symptoms 76.6%, suicidal ideation 76.5%, insomnia early night 71.4%, somatic anxiety 71.2%, insomnia late night 66.8%, diurnal variation 66.5%, loss of appetite 64.4%, psychomotor retardation 55.6%, agitation 49%, loss of weight 48.6%, hypochondriasis 39%, depersonalization 30.9%, paranoid ideas 21.5%, lack of illness insight 18%, and compulsiveness 10.1%.27 The distinct pattern or cluster of symptoms that comes to the fore in an individual could lead the way toward finding a more individualized treatment regimen, which might speed up the time to response. For example, in patients with anxiety symptoms, a combination of an antidepressant with a benzodiazepine may lead to a quick reduction in the sensation of fear. Also, sleep disturbances can be addressed specifically with hypnotic comedication or a sedating antidepressant.

Figure 2
Figure 2. Functional and structural model of different
depressive subtypes. Abbreviations: DA, dopamine; NE, norepinephrine; 5-HT,
5-hydroxy-trytophan/serotonin. After reference 24: Malhi GS, Parker GB, Greenwood J.
Structural and functional models of depression: from subtypes
to substrates. Acta Psychiatr Scand. 2005;111:94-105.
Copyright © 2004, Blackwell Munksgaard.

What symptoms are predictive
of remission and what is the timing
of their alleviation?

Despite the fact that response and remission are gaining more and more acceptance in the research community as major outcome criteria in clinical trials as a result of their important clinical implications compared with mean-score differences, most trials still rely solely on end point differences between drug and placebo for remission or response rates. Only a few report both rates, and even fewer also report the time course of response and remission rates.
Generally speaking, the course of remission and response under treatment seems to follow a typical pattern: the strongest effects in terms of a mean score reduction usually appear at the beginning of the antidepressant therapy regimen, not only in phase 3 studies sponsored by pharmaceutical companies but also in naturalistic studies. By far the most pronounced HAM-D mean score decline can thus be seen after the first 2 weeks of a newly initiated treatment. Thus in the aforementioned naturalistic trial involving 1014 inpatients, a 34% reduction in the HAM-D21 score and a 33% reduction in the Montgomery–Åsberg Depression Rating Scale (MADRS) score was evident after week 2.17 This is in excellent accordance with data from another naturalistic trial showing a 30% MADRS reduction after 2 weeks in 11 760 outpatients with major depressive disorder (MDD) treated with escitalopram.28 Clearly, nonspecific factors such as the inpatient setting or professional attention received by patients within clinical trials, and also fast-acting comedication like benzodiazepines, contribute to such fast symptom relief and support the intrinsic effect of antidepressants.

Time course of response versus remission
The time courses of remission rates versus response rates reveal a contrasting picture. In the first step of STAR*D, an investigation of citalopram over 14 weeks in 2876 outpatients with depression, the response rate was approximately 17% at week 2; subsequently, the highest response rate was at week 4 with 25%, with a constant decline thereafter to study end to about 7%.18 Remission rates started at about 11% after week 2 and went up to about 22% at week 6 and declined again to 14% at study end. Interestingly, until week 4, there were higher response than remission rates, and from week 6 onward, the relationship was the opposite, with higher remission than response rates (Figure 3).18 With respect to the latter phenomenon, a strikingly similar picture is revealed by our own data on naturalistic-treated inpatients (Figure 4, page 122).17,18 At week 6, the remission rate surpasses the response rate up to the final visit. This clearly justifies and underlines the importance of differentiating and calculating both response and remission rates as two different outcome variables. These results further imply that response usually precedes remission, which is also strengthened by the notion that achievement of response is highly predictive of remission.

Figure 3
Figure 3. Remission and response rates from Sequenced Treatment Alternatives to Relieve Depression (STAR*D) Level 1. QUID-SR, Quick Inventory of Depressive
Symptomatology—Clinician Rated.
After reference 18: Trivedi MH, Rush AJ, Wisniewski SR, et al. Evaluation of outcomes with citalopram
for depression using measurement-based care in STAR*D: implications for clinical practice. Am J Psychiatry.
2006;163:28-40. Copyright © 2006, American Psychiatric Association.

Achievement of remission is highly predictive of good functional recovery and lower relapse rates. Thus responding to an antidepressant therapy might represent advancement of the healing process in major depression. Stassen and coworkers analyzed data from 2848 patients included in double-blind randomized controlled trials, and found no difference between active drug and placebo regarding the time of onset of improvement, which was equally distributed in the placebo and active drug groups. However, significant interindividual differences in response patterns independent of the treatment condition occurred. In addition, significant differences between the active group and the placebo group could be seen in the total percentage of improvers. The authors conclude that affectively ill patients are likely to possess a common biological “resilience-like component” that largely controls recovery and can best be studied in placebo-treated patients. Active drugs might induce this natural resilience mechanism.29

Figure 4
Figure 4. Remission and response rates from a German naturalistic trial.
Based on data from reference 17.

Timing of onset of antidepressant action
Traditionally, it has been thought that standard antidepressants take about 1 month for their action to fully unfold, and that they have a delayed onset of action of at least 2 weeks. Identification of the exact timing of the onset of antidepressant action has been a well-debated matter for the last 20 years. Originally, Frederic Quitkin proposed in his pattern analytic approach that real drug-placebo differences could not be observed before 3 weeks of treatment.30 An earlier improvement was associated with a placebo response and a lack of sustained improvement, whereas the opposite was true for true drug responders, who showed a delayed but sustained onset of improvement.30,31 More recently, this view has been questioned by a large number of authors, including Stassen and Szegedi, who have not only emphasized that an earlier onset of improvement before 2 weeks was highly prevalent, but also showed that it was highly predictive of later outcome.15,32-34 An early improvement was therefore defined as a 20% reduction in the initial HAM-D17 score within the first 2 weeks. Today, there is strong evidence that early onset of antidepressant action exists and appears to be highly sensitive for later response or remission. These results can best be translated into clinical practice as follows: if after 2 weeks no improvement at all can be observed under a new antidepressant regimen, then the pharmacologic regimen should be adjusted or changed early rather than waiting for another 2 or 3 weeks. In this case, if no alteration of the treatment has occurred, the chances of this person still responding are below 20%.34
However, there are a number of methodological pitfalls inherent in the study of this phenomenon. The rating scales currently in use such as HAM-D and MADRS were never developed to detect early changes in psychopathology due to antidepressant action. In addition, most publications on this topic rely on post-hoc analysis of phase 3 trials not designed for the detection of an early onset of action, with 2-week rating intervals not allowing the detection of earlier improvements. The inconsistencies in Quitkin’s conclusions can best be explained by methodological differences between the two approaches. Quitkin et al defined stable drug response very strictly as “much improvement” on the clinical global impression scale. As pointed out by Gelenberg, it might be that the lag in the onset of antidepressant action observed by Quitkin was caused by the high degree of improvement necessary for response. 10 In other words, authors such as Stassen and Szegedi observed earlier effects, because they used lower and more sensitive cut-off values, which enabled them to detect minor changes earlier.
One of the few studies specifically designed to study the onset of action of antidepressants was conducted by Rojo and coworkers. They investigated the onset of antidepressant action in 582 depressed outpatients with HAM-D17 assessments at baseline and after 1, 2, and 4 weeks.35 Patients were treated with mirtazapine and, depending on the timing of response, sustained responders were subdivided into four groups: very fast responders, fast responders, traditional responders, and nonresponders. At baseline, only a few symptoms captured by HAM-D17 showed significant differences between the responder groups; ie, items such as guilt, suicide, and genital symptoms: the more severe these baseline variables were, the slower was the response.36

Order and timing of symptom alleviation
The order and timing of the alleviation of depressive symptoms has also rarely been studied. With respect to the Rojo study, across all four subgroups, the pattern on a symptom level did not vary: first, all groups improved on items such as depressed mood, early insomnia, and psychic anxiety. Second was improvement on the HAM-D item “general somatic symptoms.” Interestingly, this improvement pattern occurred for the very fast group within the first week, for the fast group within the second week, and for the traditional group within the third week. In summary, the alleviation of HAM-D17 symptoms for each subtype was similar, but occurred at a different time in each responder group.36 These results suggest that this equal response may not depend on personality traits, doctor-patient relationships, or medical practice, but rather on pharmacological factors and pharmacokinetics, at least in the case of mirtazapine.
In our own naturalistic follow-up study on 1014 depressed inpatients, we observed the long-term alleviation of the 6 core items of HAM-D (Seemüller et al, unpublished data). Apart from item 1, all of the other 5 items6,10,11,23,35 paralleled the same decline pattern until discharge, with the most pronounced difference between baseline and week 2. Item 1 of HAM-D17, however, referred to as “depressive feelings,” showed the strongest and fastest decline of all items up to discharge (Figure 5). In another large naturalistic trial on escitalopram in 11760 outpatients, patients were asked to name the MADRS symptoms that changed the most 2 weeks after starting antidepressant therapy. Patients most frequently answered that “inner tension,” followed by “sadness,” “pessimistic thoughts,” and “reduced sleep” had changed as early as week 2.28
Unfortunately, a substantial proportion of patients do not respond to treatment within this short time period, and most depressed patients do not initially achieve remission. In STAR*D, only 28% of the initial sample of 2876 depressed patients achieved remission, and 47% were responders with citalopram.18 The longer the search for an effective antidepressant treatment in an individual, the higher the potential risk of chronicity, the higher the overall costs, and the worse the functional outcome. If a patient could be detected and identified early on as being a nonresponder to an antidepressant and could consequently be referred earlier to a more effective strategy such as dual-acting antidepressants, augmentation strategies like lithium augmentation or augmentation with an atypical antipsychotic, or to other highly effective treatments such as ECT, the difficult challenge of optimizing the treatment of depression could be facilitated.

Clinical predictors of response to antidepressant medication
One way to classify predictors is to distinguish between biological and clinical predictors, meaning either predictors that can be measured with a distinct laboratory test or predictors that can simply be observed or asked about. In clinical research, in addition to the classification of subtypes of depression, clinical/anamnestic predictors have been examined regarding the response to antidepressants. The following have been relatively consistently described as being relevant to a rather poor response to antidepressants37,38—most have been also confirmed in later examinations18,39,40: poor social adaptation, neurotic traits in the premorbid personality, number and duration of earlier psychiatric inpatient treatments, nonresponse to earlier treatments with antidepressants, chronicity of the depressive symptoms, mild degree of depressive symptoms, delusions, and absence of vital symptoms.
However, the variance rates described by single predictors are for the most part so minor that they are hardly useable for practicable prognostics in individuals. In addition, there are hardly any hypotheses- free approaches investigating a whole set of different psychopathologic measures and clinical variables with respect to their prognostic meaning. So far, the predictors we know of can at best con- tribute to group statistical differentiation. Future research should also focus on the possibility of combining predictors, thereby potentially enabling prognosis to be optimized.

Why are available antidepressants
not always able to achieve remission?

Partial response and nonresponse to antidepressant medications are a common problem in patients with depression. Between 10% and 30% of depressed patients taking an antidepressant are partially or totally resistant to the treatment. Recurrence of depression while still taking medication (ie, breakthrough) can also occur. In our own cohort of 1014 depressed inpatients, about 32% of all patients suffered from an index episode that lasted longer then 6 months and were therefore classified as partially refractory patients.17

Figure 5
Figure 5. Individual item changes for the 6 core symptoms on the 17-item Hamilton Rating Scale
for Depression in naturalistic-treated depressed patients.

When speaking of resistance, we should clearly separate true drug resistance from what we call “pseudoresistance.” Often, patients who are labeled “treatment resistant” have not had a single adequate medication trial (either in terms of dose or duration). Another often overlooked reason for pseudoresistance is misdiagnosis, and with it goes the tendency for many psychiatrists to stick to their original diagnosis instead of reevaluating symptoms carefully. One of the most important reasons for pseudoresistance might be poor adherence. Particularly in outpatients, adherence rates can be as low as 50%, with the other 50% of patients not taking their medication properly.41 In most cases, side effects account for the poor compliance rates. Olin and colleagues summarized the results of a survey of 1066 psychiatrists conducted at the 2007 American Psychiatric Association meeting. According to the survey, about one third of the respondents considered current antidepressants to be only “somewhat effective,” and ranked sexual dysfunction, weight gain, and inadequate efficacy as the primary roadblocks to successful treatment.
In addition, comorbidities such as substance abuse, personality disorders, and general medical disorders including hypothyroidism and anemia can also influence the overall success of a treatment plan and can muddy the water. In particular, medical comorbidities might negatively influence outcome in two ways: first, through the direct impact of the medical comorbidity on the psychiatric illness, and second, through adverse effects of medications used to treat the medical comorbidities, also known as pharmacogenic depression. With regard to relapse rates after complete remission, Andrew Nierenberg demonstrated most recently in his talk at the Collegium Internationale Neuro-Psychopharmacologicum congress in Munich, showing data from STAR*D, that the higher the load for psychiatric comorbidities, the higher the risk of relapse over 1 year.42
The reasons for true drug resistance remain unknown. Unfortunately, even the definition of drug resistance is not fully standardized and varies greatly across different studies. In daily clinical practice, probably the most useful definition might be as follows: drug nonresponse to at least two adequate trials of antidepressants (adequate dosage and treatment duration) with different mechanisms of action. Given the aforementioned complexity and diversity of the etiopathogenesis of such a multifaceted illness like depression, a “one treatment fits all” approach to regaining a new homeostasis on a transmitter level and a functional level at the same time cannot be very fruitful.
New ideas regarding the underlying pharmacological mechanisms that cause a poor response may lead to a more personalized therapy that is based on “the right drug for the right patient.” One new way forward might arise from a recent study on the ABCB1 drug transporter gene, in which the basic idea was that every antidepressant drug must pass the blood brain barrier before its antidepressant ac- tivity can unfold. P-glycoprotein (P-gp),43 which is encoded on the ABCB1gene, serves as a transporter and acts as an efflux pump for a wide range of psychopharmacological compounds. Uhr and colleagues were able to identify the antidepressant substrates of P-gp, and could predict response and remission in relation to the genotype of subjects who were treated with antidepressants that were substrates of P-gp.43 A completely different approach may arise from the use of brain imaging techniques. Frodl and coworkers, for example, were able to demonstrate that nonresponders to antidepressant treatment showed significantly lower amygdala volumes at baseline than did remitters.35


In summary, aside from the complexity of this inhomogeneous illness, several factors may contribute to treatment failure, including undiagnosed or misdiagnosed medical conditions such as hypothyroidism and anemia. The patient who does not respond or only partially responds to an antidepressant should therefore first be reassessed to make sure the original diagnosis of depression was correct. Also, nonpsychiatric drugs such as methyldopa, â-blockers, and reserpine can cause or exacerbate depression, and comorbid disorders (eg, eating disorders, substance abuse or dependence) may affect treatment response. Psychotic depression, melancholic depression, and atypical depression are depressive subtypes that may require concurrent pharmacotherapy such as antipsychotic or augmentative psychotherapy. Finally, adverse effects and poor compliance may be additional obstacles to successful treatment, but with our growing understanding of the pathophysiology of the illness, there is hope that in the near future we will be making progress in the direction of individualizing the clinical decision-making process.❒


1. Kendler KS. Explanatory models for psychiatric illness. Am J Psychiatry. 2008;165:695-702.
2. Belmaker RH, Agam G. Major depressive disorder. N Engl J Med. 2008;358:55-68.
3. Kendler KS, Karkowski LM, Prescott CA. Causal relationship between stressful life events and the onset of major depression. Am J Psychiatry. 1999;156:837-841.
4. Caspi A, Sugden K, Moffitt TE, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science. 2003;301:386-389.
5. Ruhe HG, Mason NS, Schene AH. Mood is indirectly related to serotonin, norepinephrine and dopamine levels in humans: a meta-analysis of monoamine depletion studies. Mol Psychiatry. 2007;12:331-359.
6. Burke HM, Davis MC, Otte C, Mohr DC. Depression and cortisol responses to psychological stress: a meta-analysis. Psychoneuroendocrinology. 2005;30:846-856.
7. Merali Z, Du L, Hrdina P, et al. Dysregulation in the suicide brain: mRNA expression of corticotropin-releasing hormone receptors and GABA(A) receptor subunits in frontal cortical brain region. J Neurosci. 2004;24:1478-1485.
8. Carroll BJ, Cassidy F, Naftolowitz D, et al. Pathophysiology of hypercortisolism in depression. Acta Psychiatr Scand. 2007(suppl); 90-103.
9. Flores BH, Kenna H, Keller J, Solvason HB, Schatzberg AF. Clinical and biological effects of mifepristone treatment for psychotic depression. Neuropsychopharmacology. 2006;31:628-636.
10. Frodl T, Schule C, Schmitt G, et al. Association of the brainderived neurotrophic factor Val66Met polymorphism with reduced hippocampal volumes in major depression. Arch Gen Psychiatry. 2007;64:410-416.
11. Karege F, Vaudan G, Schwald M, Perroud N, La Harpe R. Neurotrophin levels in postmortem brains of suicide victims and the effects of antemortem diagnosis and psychotropic drugs. Brain Res Mol Brain Res. 2005;136:29-37.
12. Tang SW, Chu E, Hui T, Helmeste D, Law C. Influence of exercise on serum brain-derived neurotrophic factor concentrations in healthy human subjects. Neurosci Lett. 2008;431:62-65.
13. Marano CM, Phatak P, Vemulapalli UR, et al. Increased plasma concentration of brain-derived neurotrophic factor with electroconvulsive therapy: a pilot study in patients with major depression. J Clin Psychiatry. 2007;68:512-517.
14. Hellweg R, Ziegenhorn A, Heuser I, Deuschle M. Serum concentrations of nerve growth factor and brain-derived neurotrophic factor in depressed patients before and after antidepressant treatment. Pharmacopsychiatry. 2008;41:66-71.
15. Kennedy SH. Agomelatine: an antidepressant with a novel mechanism of action. Future Neurology. 2007;2:145-151.
16. Nierenberg AA, McLean NE, Alpert JE, Worthington JJ, Rosenbaum JF, Fava M. Early nonresponse to fluoxetine as a predictor of poor 8-week outcome. Am J Psychiatry. 1995;152:1500-1503.
17. Seemuller F, Riedel M, Obermeier M, et al. Outcomes of 1014 naturalistic treated inpatients with major depressive episode. Poster presented at: XXVI CINP Congress; July 13-17, 2008; Munich, Germany.
18. Trivedi MH, Rush AJ, Wisniewski SR, et al. Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. Am J Psychiatry. 2006;163:28-40.
19. Hagerty BM, Williams RA, Liken M. Prodromal symptoms of recurrent major depressive episodes: a qualitative analysis. Am J Orthopsychiatry. 1997;67:308-314.
20. Perlis ML, Giles DE, Buysse DJ, Tu X, Kupfer DJ. Self-reported sleep disturbance as a prodromal symptom in recurrent depression. J Affect Disord. 1997;42:209-212.
21. Roper PD. Prodromal symptoms. Am J Psychiatry. 1989;146: 941-942.
22. Dryman A, Eaton WW. Affective symptoms associated with the onset of major depression in the community: findings from the US National Institute of Mental Health Epidemiologic Catchment Area Program. Acta Psychiatr Scand. 1991;84:1-5.
23. Hegerl U, Bottner AC, Holtschmidt-Taschner B, et al. Onset of depressive episodes is faster in patients with bipolar versus unipolar depressive disorder: evidence from a retrospective comparative study. J Clin Psychiatry. 2008;69:1075-1080.
24. Malhi GS, Parker GB, Greenwood J. Structural and functional models of depression: from subtypes to substrates. Acta Psychiatr Scand. 2005;111:94-105.
25. Henkel V, Seemuller F, Obermeier M, et al. Does early improvement of depressive syndromes triggered by antidepressant compounds predict remission? Analysis of data of a naturalistic study on a large sample of depressed inpatients. World J Biol Psychiatry. In press.
26. Seemüller F, Riedel M, Obermeier M, et al. The controversial link between antidepressants and suicidality risks in adults: do data from the German Depression Katamnesis study support the FDA decisions of suicide warnings for antidepressants? Eur Arch Psychiatry Clin Neurosci. In press.
27. Seemuller F, Riedel M, Wickelmaier F, et al. Atypical symptoms in hospitalised patients with major depressive episode: frequency, clinical characteristics, and internal validity. J Affect Disord. 2008;108:261-278.
28. Möller HJ, Langer S, Schmauss M. Escitalopram in clinical practice: results of an open-label trial in outpatients with depression in a naturalistic setting in Germany. Pharmacopsychiatry. 2007;40:53-57.
29. Stassen HH, Angst J, Hell D, Scharfetter C, Szegedi A. Is there a common resilience mechanism underlying antidepressant drug response? Evidence from 2848 patients. J Clin Psychiatry. 2007; 68:1195-1205.
30. Quitkin FM, Rabkin JD, Markowitz JM, Stewart JW, McGrath PJ, Harrison W. Use of pattern analysis to identify true drug response. A replication. Arch Gen Psychiatry. 1987;44:259-264.
31. Quitkin FM, McGrath PJ, Stewart JW, Taylor BP, Klein DF. Can the effects of antidepressants be observed in the first two weeks of treatment? Neuropsychopharmacology. 1996;15:390-394.
32. Nierenberg AA, Farabaugh AH, Alpert JE, et al. Timing of onset of antidepressant response with fluoxetine treatment. Am J Psychiatry. 2000;157:1423-1428.
33. Stassen HH, Angst J, Delini-Stula A. Delayed onset of action of antidepressant drugs? Survey of results of Zurich meta-analyses. Pharmacopsychiatry. 1996;29:87-96.
34. Szegedi A, Muller MJ, Anghelescu I, Klawe C, Kohnen R, Benkert O. Early improvement under mirtazapine and paroxetine predicts later stable response and remission with high sensitivity in patients with major depression. J Clin Psychiatry. 2003; 64:413-420.
35. Frodl T, Meisenzahl EM, Zetzsche T, et al. Hippocampal and amygdala changes in patients with major depressive disorder and healthy controls during a 1-year follow-up. J Clin Psychiatry. 2004;65:492-499.
36. Rojo JE, Gibert K, Cobo J, Rodriguez-Cano E, Vallejo J. Onset of antidepressant action: a pharmacological question? Hum Psychopharmacol. 2005;20:425-433.
37. Klein DF. Listening to meta-analysis but hearing bias. Prevention & Treatment. 1998;1:Article ID 0006c. June 26,1998.
38. Möller HJ, Fischer G, von Zerssen D. Prediction of therapeutic response in acute treatment with antidepressants. Results of an empirical study involving 159 endogenous depressive inpatients. Eur Arch Psychiatry Neurol Sci. 1987;236:349-357.
39. Seemüller F, Möller HJ, Obermeier M, et al. Prediction of response and remission in inpatients with depressive syndromes. World J Biol Psychiatry. In press.
40. Sharan P, Saxena S. Treatment-resistant depression: clinical significance, concept and management. Natl Med J India.1998; 11:69-79.
41. Linden M. Therapeutic strategies for improving patient compliance (author’s translation from German). Nervenarzt.1979;50: 109-114.
42. Gilmer WS, Gollan JK, Wisniewski SR, et al. Does the duration of index episode affect the treatment outcome of major depressive disorder? A STAR*D report. J Clin Psychiatry. 2008;69: 1246-1256.
43. Uhr M, Tontsch A, Namendorf C, et al. Polymorphisms in the drug transporter gene ABCB1 predict antidepressant treatment response in depression. Neuron. 2008;57:203-209.


La dépression majeure est une maladie complexe et hétérogène dont l’étiopathogenèse est basée sur de nombreux facteurs qui peuvent agir à différents niveaux (psychologique, biologique, génétique et social). Il est donc avéré qu’un traitement unique ou une approche thérapeutique unique ne peut convenir à une maladie aussi diversifiée. Nous avons l’espoir dans un futur proche de pouvoir traiter les patients de façon encore plus personnalisée grâce aux marqueurs cérébraux morphologiques et génétiques. En cas de résistance au traitement, une pseudorésistance (en insistant sur l’observance) doit être tout d’abord éliminée. La réponse au traitement peut être très précoce, dès les 2 premières semaines du traitement, mais dans les études cliniques le taux le plus élevé de réponses se situe autour des 4 premières semaines. Quand exactement et dans quelle mesure la réponse apparaît, dépendent aussi de sa définition exacte. La communauté scientifique a adopté un consensus sur deux définitions : celle de l’amélioration précoce (20 % de réduction sur l’échelle de HAM-D [Hamilton Rating Scale for Depression] au cours des 2 premières semaines) et celle de la réponse proprement dite (_ 50 % de réduction du score initial HAM-D). La première est surtout utile dans la surveillance du soulagement des symptômes grâce au traitement en question alors que la seconde est plus utilisée pour la détection de modifications prolongées, particulièrement en association avec un critère temporel. L’amélioration précoce et la réponse, ainsi définies, sont fortement prédictives de rémission. Les études cliniques indiquent que si les taux de rémission sont habituellement plus faibles que ceux de réponse, ils les dépassent néanmoins à partir de la 4e à la 6e semaine. La réponse au traitement apparaît donc comme un précurseur de la rémission, et donc comme la phase initiale de l’évolution naturelle de la dépression sous traitements antidépresseurs, qui dans la majorité des cas aboutira à la rémission, c’est à dire à la disparition complète de tous les symptômes.