Controversal question: Is early improvement predictive of antidepressant response?



Back to summary |

1. E. Aguglia and A. Petralia,Italy
2. Y-M. Ahn,Korea
3. A. S. Avedisova,Russia
4. I. Bitter,Hungary
5. M. L. Figueira,Portugal
6. S. Kuasirikul,Thailand
7. G. Parker,Australia
8. M. Roca,Spain
9. S. Vahip,Turkey
10. M. Wong,Hong Kong
11. R. M. Zaratiegui,Argentina

1. E. Aguglia and A. Petralia,Italy

Eugenio AGUGLIA, MD
Full Professor of Psychiatry
Antonino PETRALIA, MD
School of Medicine, University of Catania
S. Sofia 8, 95123 Catania, ITALY
(e-mail: e.aguglia@live.it)

A fundamental issue in antidepressant therapy concerns predictors of therapeutic response. There is numerous evidence in the literature1-4 showing that the delayed appearance of a therapeutic effect, in addition to increasing the risk of suicide, prolongs the individual’s and family’s suffering, while an early improvement increases positive outcomes and the compliance with treatment. Szegedi et al1 suggest that improvement in an individual in the early stage of treatment is a predictor of future response.

Previously, Segman et al2 and more recently, van Calker et al3 reported that early improvement predicts later response, remission, and resistance to treatment, because in their studies, those who didn’t show an early improvement (within the second week of treatment) had a low probability of achieving a delayed therapeutic response or remission. Some authors have stated that clinical improvement within the second week of treatment predicts, with high sensibility, the responses in the 4th and 6th weeks.4 Similar results were obtained in a recent naturalistic prospective study conducted on a sample of 795 patients with major depression,5 in which it was shown that an improvement in the first 2 weeks of treatment predicted the rate of delayed response and remission, even in hospitalized patients with more serious disease.

According to researchers, a low or no initial response to treatment could justify an accelerated switch to alternative treatment. There are many variables that directly influence the latency time to clinical improvement. Several studies have shown population characteristics that may be identified as possible factors affecting the latency time to therapeutic response, such as the episode’s depression severity, the duration of the episode before the start of treatment, and the medical history of response to previous drug treatments.6 These aspects, together with the pharmacodynamic and pharmacokinetic characteristics of the drug, as well as characteristics of the individual patient, such as age, personality, and genetic variables (T and C variants of the 5-HT2A receptor), contribute to modify the said latency time.

Other research has indicated some variables involved in a poor therapeutic response, such as low socioeconomic status and the presence of anxiety symptoms in the period preceding the start of treatment. These variables in elderly depressed patients are responsible for increased suicidal ideation. Another study identified six patient characteristics, in addition to depression severity, that are involved in the outcome of various antidepressant treatment types: social dysfunction, cognitive dysfunction, expectation of improvement, endogenous depression, double depression, and the duration of current episode.

Therefore, we must take into account all those factors that can increase the latency period and hence the disease time, thus avoiding the absence of remission, presence of residual symptoms, or resistance to treatment, because an early response to antidepressant therapy implies a better result for the entire duration of treatment, thus constituting a factor that is predictive of results in the medium and long term. _

References
1. Szegedi A, Jansen WT, van Willigenburg AP, van der Meulen E, Stassen HH, Thase ME. Early improvement in the first 2 weeks as a predictor of treatment outcome in patients with major depressive disorder: a meta-analysis including 6562 patients. J Clin Psychiatry. 2009;70:344-353.
2. Segman RH, Shapira B, Gorfine M, Lerer B. Onset and time course of antidepressant action: psychopharmacological implications of a controlled trial of electroconvulsive therapy. Psychopharmacology (Berl). 1995;119:440-448.
3. van Calker D, Zobel I, Dykierek P, et al. Time course of response to antidepressants: predictive value of early improvement and effect of additional psychotherapy. J Affect Disord. 2009;114:243-253.
4. Kennedy SH, Eisfeld BS, Meyer JH, Bagby RM. Antidepressants in clinical practice: limitations of assessment methods and drug response. Hum Psychopharmacol. 2001;16:105-114.
5. Henkel V, Seemüller F, Obermeier M, et al. Does early improvement triggered by antidepressants predict response/remission? Analysis of data from a naturalistic study on a large sample of inpatients with major depression. J Affect Disord. 2009;115:439-449.
6. Hennings JM, Owashi T, Binder EB, et al. Clinical characteristics and treatment outcome in a representative sample of depressed inpatients—findings from the Munich Antidepressant Response Signature (MARS) project. J Psychiatr Res. 2009;43:215-229.

2. Y-M. Ahn,Korea

Yong Min AHN, MD, PhD
Department of Neuropsychiatry
Seoul National University College of Medicine
Seoul National University Hospital
28 Yeongeon-Dong, Jongno-Gu
Seoul, KOREA
(e-mail: aym@snu.ac.kr)

In the treatment of depression, clinicians frequently need to carry out early identification of nonresponders and promptly implement an alternative treatment strategy deemed to be superior, rather than waiting 4-6 weeks on antidepressant treatment, as stated by the current guidelines. Such practice invariably serves to reduce suicide risk and unnecessary drug exposure during ineffective treatment, while increasing adherence and the chance of a better outcome.1

Current guidelines are based on results indicating that the antidepressant effects within the first few weeks are in reality likely to be a placebo response of an abrupt and nonpersistent nature, dissimilar to a true drug effect.1,2 This evidence of a delayed response to antidepressants has been widely accepted in psychiatric practice and research fields. However, data regarding early onset of antidepressant action within the first 14 days of treatment have recently increased. In addition, an early therapeutic effect has been suggested to be the best predictor of response to antidepressant at end point,3,4 and is also positively related to the restoration of psychosocial functioning.5 A recent large meta-analysis of 41 clinical trials including 6562 depressive patients, showed that early improvement at 2 weeks predicted response and stable remission, with high sensitivity of above 80%. In this meta-analysis, negative predictive values for stable response and remission were very high (82%-100%), but positive predictive value was relatively low (19%-60%).1

When applying these data regarding early improvement as a response predictor in real clinical practice, it is necessary to consider the limitations of the data, as most studies were not specifically designed for such a purpose. In addition, thorough examination of factors such as the cut-off values for early improvement, later response, and remission, and the decision time point is necessary. The definitions of such factors were made somewhat arbitrarily, and are not suitable for real practice. Studies have not yet yielded any information regarding the long-term outcome (eg, recurrences) and the differences among many kinds of antidepressants (eg, selective serotonin reuptake inhibitors [SSRIs] vs serotonin norepinephrine reuptake inhibitors).

Now, I’d like to address the reported possibility of there being differential improvement patterns for several depressive symptoms.6 All symptoms of depression did not improve simultaneously. The order and degree of improvement in depressive symptomatology might result from different neurochemical actions among antidepressants. Certain symptoms may improve more quickly than others (eg, anxiety and depressive mood with SSRIs vs psychomotor retardation with desipramine).4 In such cases, a symptom-specific measure may be more sensitive than measurement of global symptom severity for detecting early antidepressant effects.

It remains uncertain as to which type of symptom improvement would be more predictive of long-term outcome. The emotional domain (eg, depressive mood and anhedonia) has been regarded as the classical core of depression. Recently, however, anxiety, residual somatic symptoms, and cognitive dysfunction have increasingly received attention as treatment targets. Early improvement of anxiety might predict higher response and remission. Also, it is known that residual sleep disturbance and sexual dysfunction are significantly associated with subsequent relapse or recurrence and subsequent poor treatment outcome. The biological bases of these symptoms in depression were studied extensively in the last decade.

Despite many limitations and problems, in light of consistently high negative predictive values obtained in recent data, it would be advisable to alter the current treatment strategy in the case of insufficient improvement within a couple of weeks of standard antidepressant treatment. Prospective studies specifically designed to address this issue are needed in the future. _

References
1. Szegedi A, Müller 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.
2. Quitkin FM, Rabkin JG, Ross D, Stewart JW. Identification of true drug response to antidepressants. Use of pattern analysis. Arch Gen Psychiatry. 1984;41: 782-786.
3. Henkel V, Seemüller F, Obermeier M, et al. Does early improvement triggered by antidepressants predict response/remission? Analysis of data from a naturalistic study on a large sample of inpatients with major depression. J Affect Disord. 2009;115:439-449.
4. Katz MM, Tekell JL, Bowden CL, et al. Onset and early behavioral effects of pharmacologically different antidepressants and placebo in depression. Neuropsychopharmacology. 2004;29:566-579.
5. Papakostas GI, Petersen T, Denninger JW, et al. Psychosocial functioning during the treatment of major depressive disorder with fluoxetine. J Clin Psychopharmacol. 2004;24:507-511.
6. Boyer P, Tassin JP, Falissart B, Troy S. Sequential improvement of anxiety, depression and anhedonia with sertraline treatment in patients with major depression. J Clin Pharm Ther. 2000;25:363-371.

3. A. S. Avedisova,Russia

Alla S. AVEDISOVA, MD
Department of New Drugs and
Methods Therapy
Serbsky National Research Centre
for Social and Forensic Psychiatry
Moscow, RUSSIA
(e-mail: aavedisova@hotmail.com)

A number of studies on efficiency predictors in the treatment of major depressive disorder have been based on the standardization of treatment efficiency evaluation, using remission (RM), response (RS), partial response (PR), and nonresponse (NR) criteria. In one study involving 130 patients, researchers studied “baseline” predictors (social and clinical characteristics at beginning of treatment) and “process” predictors (changes during the 8-week antidepressant treatment course, including time to initial response [IRS], ie, a reduction of at least 25% on the Hamilton Rating Scale for Depression [HAM-D]). Stratification of various drug treatment efficiency predictors showed IRS time to be the most important. A total of 24.6% of studied patients developed IRS during the first week of treatment. Of these, 68.8% were categorized as being in RM, and 31.2% as RS. Patients with longer IRS times less frequently achieved RM and RS and more frequently achieved PR. For patients who developed IRS during the 2nd treatment week, 46.2% achieved RM, 36.5% RS, and 6.9% PR. For patients with IRS in the 3rd treatment week, 28.6% achieved RM, 34.3% RS, and 37.1% PR. For patients with IRS in the 4th week, 2.3% achieved RS and 17.3% PR. For patients with IRS in the 5th week (26.9% of all patients), 33.3% achieved RS and 66.7% PR. Two patients developed IRS in the 6th or the 7th week. One of these achieved RS, the other PR. None of the nonresponders developed an IRS. This regularity was found to be independent of the antidepressant administered.

The period during which IRS develops is crucial, as it determines the efficiency of later treatment stages. However, this is a “silent” (latent) period regarding antidepressant effect. Results indicate the absence of a strong correlation between the development of IRS and the antidepressant action of the drug. At the same time, this period is remarkable for its most evident placebo effect. The role of this period in the prediction of antidepressant treatment efficiency was studied in 83 patients, who underwent 1 week of placebo treatment followed by 4 weeks of antidepressants. A total of 46.9% of all major depressive disorder patients (HAM-D, 22.8 ±3.7) became placebo responders during the first week of placebo treatment. By the end of antidepressant therapy, all of them entered the RS group (HAM-D score reduction of _50%). As for placebo-NR patients (53.1% of all participants), only 55% of them became responsive to antidepressant therapy. The results of this study indicate that response to placebo is a complex reaction of the whole human body, including multiple changes at the clinical, neurochemical, and neurophysiological levels. The monoamine excretion test revealed increased activity of the catecholamine neurotransmitter system in the placebo-responsive group (n=15) compared with placebo nonresponders (n=18), who did not develop the same changes (placebo responders: histamine background level 14.8 ±1.6 ng/min, placebo period 18.9 ±1 ng/min [P<0.01]; dopamine background level 112.6 ±2.3 ng/min, placebo period 124.8 ±2.6 ng/min). The same was true for the indolamine system (placebo responders: 5-OT background level 116.4 ±2.6 ng/min, placebo period 127.8 ±1.9 ng/min [P<0.05]). Test results confirmed the neurophysiological basis of the placebo effect (narrow-band spectrum analysis of multichannel electroencephalogram was used). This effect features a marked single-type reaction of an increase in the low and medium-frequency range of á-rhythm (6.3-9.0 Hz and 9.0- 10.2 Hz), in contrast with nontypical, feebly marked single bands (mostly of â1-rhythm) in placebo-NR (P<0.05). These data demonstrate higher neurophysiological reactivity in placebo- responsive patients. Further antidepressant therapy was found to lead to a completely different electroencephalogram profile (restoration of physiological asymmetry in low- and medium-frequency á-rhythm signals in central and frontal areas of the brain). These changes developed only in patients who were responsive to active therapy. These data support the idea of the initial reactivity of neurobiological mechanisms to certain drugs in placebo-RS patients (which may be genetically determined). The manifestations in the early stages of therapy include sensitivity to placebo and to nonspecific actions of antidepressants. This may also be related to the neuroplasticity of certain cerebral structures (brainderived neurotrophic factor hypothesis). IRS may be considered to be a specific independent predictor of RMachievement. The new major depressive disorder treatment strategy implies a personalized approach to the adjustment of the treatment regimen, dependent on a patient’s individual IRS time. _

4. I. Bitter,Hungary

Istvan BITTER, MD, PhD, DSc
Professor and Chair
Department of Psychiatry
and Psychotherapy
Semmelweis University
Balassa u.6, 1083 Budapest
HUNGARY
(e-mail: bitter@psych.sote.hu)

Time to onset of action and time to response are of great clinical importance; however, they have rarely been included as an outcome variable in older depression studies. While there are patients who have an early response to antidepressant treatment, most antidepressant treatments suffer from limited efficacy and a slow onset of action. Slow, delayed onset of action and delayed response during antidepressive treatment—which has already been described in the classical textbooks—is a source of frustration for patients and their families, a major reason for nonadherence, and may contribute to suicidality.

The predictive value of early improvement and response or nonresponse has been investigated. Szegedi et al1 published data suggesting that early improvement predicts later stable response and remission with high sensitivity. Improvement occurred in 65% of paroxetine-treated patients and 73% of mirtazapine- treated patients within 2 weeks. On the other hand, early nonresponse may predict poor later outcome: nonresponse to fluoxetine predicted the 8-week outcome as early as week 2.2 The results of a meta-analysis suggest that “true” antidepressant response can occur in the first 2 weeks as well as the first week of treatment of major depressive disorder.3 A recent naturalistic study on a large sample (n=795) of inpatients with major depression confirmed the findings of randomized controlled trials: early improvement in the first 2 weeksmay predict later response and remission with high sensitivity, even in hospitalized patients suffering from a more severe degree of depression.4 Fewer side effects and early onset of action both contribute to an improved adherence to medication.

Thus it is important to develop antidepressant drugs that have an early onset of action. Drugs with a dual action demonstrated this benefit in several studies; for example, in two studies, a significantly greater proportion of patients treated with venlafaxine than placebo had a clinically meaningful drug response within the first 2 weeks of treatment, and this early response persisted for the duration of the studies.5 Different antidepressants may have different symptom profiles for early improvement/response; eg, desipramine treatment was found to be associated with early improvement in motor retardation and depressed mood, while anxiety and hostility symptoms showed early response to paroxetine. Agomelatine, an antidepressant with a new mode of action, has also been demonstrated to have an early onset of action, especially in the improvement of sleep, which is included in its Summary of Product Characteristics: “From the first week of treatment, onset of sleep and the quality of sleep were significantly improved without daytime clumsiness as assessed by patients.”6 A study comparing agomelatine with venlafaxine and another study comparing agomelatine with sertraline demonstrated earlier and superior improvement in wake/sleep disorders associated with depression.

Time to response data help in deciding how long to continue with a drug treatment if it has not yet shown an onset of action. Time to response data also help in making some other treatment decisions, such as dose increases or augmentation of the current treatment with another drug or psychotherapy. Early onset of action may improve adherence, predict later response and remission with high sensitivity, and also reduce relapse rates even in hospitalized patients suffering from a more severe degree of depression. In conclusion, available data support the notion that early onset of therapeutic action in depression is related to greater efficacy of antidepressant treatments. _

References
1. Szegedi A, Müller M, 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.
2. 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.
3. Papakostas GI, Perlis RH, Scalia MJ, Petersen TJ, Fava M. A meta-analysis of early sustained response rates between antidepressants and placebo for the treatment of major depressive disorder. J Clin Psychopharmacol. 2006;26:56-60.
4. Henkel V, Seemuller F, Obermeier M, et al. Does early improvement triggered by antidepressants predict response/remission? Analysis of data from a naturalistic study on a large sample of inpatients with major depression. J Affect Disord. 2009;115:439-449.
5. Entsuah R, Derivan A, Kikta D. Early onset of antidepressant action of venlafaxine: pattern analysis in intent-to-treat patients. Clinical Ther. 1998;20:517-526.
6. Valdoxan: Summary of Product Characteristics. http://www.valdoxan.com/ index.php/summary-of-product-characteristics. Accessed August 28, 2009.

5. M. L. Figueira,Portugal

Maria L. FIGUEIRA, MD, PhD
Professor, Head of Department
of Psychiatry
Santa Maria University Hospital
Lisbon, PORTUGAL
(e-mail: ml.figueira@hsm.min-saude.pt)

Several antidepressant drugs are commonly used for the treatment of this serious condition, depression, with one of the greatest pitfalls of currently-available antidepressants being their latency of therapeutic effect. Since most, if not all, of them usually require 3 to 4 weeks to achieve symptom relief, one is commonly faced with patients still experiencing symptoms, leading to functional impairment and increased suicide risk. Such latency of effect has been commonly linked to secondary psychosocial deficits.

One is faced with a clinical picture of urgent demand for prompt and efficacious treatment in order to restrict the latency time to achievement of therapeutic effect. Such an effect, despite patients taking an antidepressant for weeks, is not attained by a non-neglectable number of patients, who despite long-term antidepressant exposure, still fail to achieve remission. Roughly half of depressed patients fail to respond to the first prescribed antidepressant, while two thirds will fail to achieve remission. An absence of response within the first 2 weeks of antidepressant treatment may lead to a lower response probability at a later time frame.1 Thus, shortening the antidepressant latency, coupled with reducing both partial response and failure to achieve remission, becomes a quest in itself.

Data from controlled studies (post-hoc analyses), meta-analysis, and even large-scale observational studies in clinical settings have questioned the antidepressant latency length, suggesting that some current antidepressant treatments can exert some initial beneficial effects early, ranging from the first to the second week of the patient’s exposure.2 Although a number of methodological study limitations can be identified relating to insufficient measurement of early response and data arising from studies not specifically designed to assess speed of antidepressant action, the search for treatment efficacy predictors is utterly relevant and can be translated into an effort to identify factors involved in non or partial remission. Among these, depression severity, length of episode before treatment commenced, concomitant anxious symptoms, comorbid disorders, and painful or physical symptoms like fatigue have been reported. With regard to the latter, if present at the onset of depression treatment, there is a lower chance of achieving remission.3

So far, the most sensitive predictor of stable response and remission available is the decrease of depressive symptoms at an early stage of treatment. Such a predictor appears to be independent of the antidepressant (either monotherapy or combination), even when combined with psychotherapeutic approaches.4

Nevertheless, one has to emphasize that a fast response to an antidepressant treatment might not be due to a specific effect of the treatment, and factors like the placebo effect (especially in sudden- and fast-onset improvement) or non– treatment-specific effects also have to be considered. Differentiation of pharmacological from placebo effects is reflected in the lack of sustained improvement in the short term.5

Symptoms like depressed mood and psychomotor slowing, as well as physical symptoms such as pain, are the symptoms that through their improvement, appear to lead to remission. However, even if there is improvement in these symptoms, one should exert caution regarding residual symptoms of depression, which, if present, are predictors of future relapse and reduced psychosocial adjustment. The clinician should be aware of sleep disturbances, sexual dysfunction, fatigue, and excessive daytime sleepiness, which are the most common residual symptoms, and for which treatment should be adjusted from the very beginning.

Not yet at a clinical level, recent research has produced promising results with the use of frontal quantitative electroencephalography, which, combined with an antidepressant treatment response index, has been shown to be capable of predicting response as early as the first week with at least two antidepressants from different classes.6 _

References
1. Hennings JM, Owashi T, Binder EB, et al. Clinical characteristics and treatment outcome in a representative sample of depressed inpatients—findings from the Munich Antidepressant Response Signature (MARS) project. J Psychiatr Res. 2009;43:215-229.
2. Machado-Vieira R, Salvadore G, Luckenbaugh DA, Manji HK, Zarate CA Jr. Rapid onset of antidepressant action: a new paradigm in the research and treatment of major depression. J Clin Psychiatry. 2008;69:946-958.
3. Leuchter AF, Husain MM, Cook IA, et al. Painful physical symptoms and treatment outcome in major depressive disorder: a STAR*D (Sequenced Treatment Alternatives to Relieve Depression report. Psychol Med. 2009;3:1-13.
4. Tadi ´ c A, Helmreich I, Mergl R, et al. Early improvement is a predictor of treatment outcome in patients with mild major, minor or subsyndromal depression. J Affect Disord. 2009, May 8. Epub ahead of print.
5. Henkel V, Seemüller F, Obermeier M, et al. Does early improvement triggered by antidepressants predict response/remission? Analysis of data from a naturalistic study on a large sample of inpatients with major depression. J Affect Disord. 2009;115:439-449.
6. Leuchter AF, Cook IA, Marangell LB, et al. Comparative effectiveness of biomarkers and clinical indicators for predicting outcomes of SSRI treatment in major depressive disorder: results of the BRITE-MD study. Psychiatry Res. 2009; 169:124-131.

6. S. Kuasirikul,Thailand

Surachai KUASIRIKUL, MD
Thai Board of Psychiatry
Assistant Professor and Medical Director
Manarom Hospital and Sleep Disorders Clinic
Manarom, THAILAND
(e-mail: surachai@manarom.com)

Depressive disorders affect approximately 5% of the population each year, and are now the fourth leading cause of the global disease burden, and the leading cause of disability worldwide. Depression seriously reduces quality of life for individuals and their families, is a risk factor for suicide, and often worsens the outcome of other physical health problems.1 Current antidepressants usually require several weeks to produce beneficial clinical effects, and are only effective in achieving remission (minimal to no symptoms) in less than half of depressed patients after acute antidepressant treatment. The lessons from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, which used a 50% score reduction on the 17-item Hamilton Rating Scale for Depression (HAM-D17) to define response, confirmed that only approximately half of the depressed patients in the study really benefited or responded to the antidepressant treatment.2

Full response to antidepressant pharmacotherapy is evident only after several weeks, but considerable improvements may already be visible within the first 2 weeks.3 The definition of “early improvement” by Henkel et al is a 50% improvement of the HAM-D21 baseline total score at day 14; remission being defined as a score of ≤7 at discharge. These authors also found that a 20% reduction of the HAM-D21 baseline total score at day 14 predicts response with 75% sensitivity and 59% specificity.4

The several attempts to identify benefits of early improvement in predicting antidepressant response4 remain at present controversial, possibly due to several study limitations. First, studies such as that of Mouchabac et al looking at residual symptoms after treated major depressive disorder have found some difficulty in directly linking outcomes in the clinical setting with the HAM-D score. This particular author concluded that poor response and remission are basically due to remaining residual symptoms at the time of response, and the effect of a delay in initiating treatment, which ultimately impacts negatively on early response.5 Second, at the response onset, early behavioral effects and the clinical response to antidepressants may involve improvement in some symptoms, but not in others. This reflects only effects that are directly due to the pharmacological properties of the antidepressant and its plasma concentration, but not improvement that is due to the etiology and prognosis of depression.6 Third, early antidepressant response in some studies may be gender specific, which may be explained by different auto-endocrinology responses. Fourth, the early response to antidepressants can be attributed to pharmacogenetic susceptibility, which may involve drug metabolism and hence early onset. Last, depressed patients with insomnia tend to have poor clinical outcomes despite the early response of some of their depressive symptoms. Moreover, several studies have shown that clinical response to various antidepressant therapies can be predicted by sleep electroencephalography parameters.

In conclusion, whether early improvement is predictive of antidepressant efficacy remains to be determined. Evaluation of antidepressant efficacy needs more than just evaluation of the pattern and timing of symptom alleviation and outcome, but should include evaluation of detrimental prognostic symptoms such as sleep disturbance or insomnia as well as other residual symptoms, which need an early response for the patient to achieve full remission. _

References
1. Üstün TB, Ayuso-Mateos JL, Chatterji S, et al. Global burden of depressive disorders: methods and data sources. Br J Psychiatry. 2004;184:386-392.
2. Rush AJ, Trivedi MH, Wisniewski SR, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry. 2006;163:1905-1917.
3. Mitchell AJ. Two-week delay in onset of action of antidepressants: new evidence. Br J Psychiatry. 2006;88:105-106.
4. Henkel V, Seemüller F, Obermeier M, et al. Does early improvement triggered by antidepressants predict response/remission? Analysis of data from a naturalistic study on a large sample of inpatients with major depression. J Affect Disord. 2009;115:439-449.
5. Mouchabac S, Ferreri M, Cabanac F, Bitton M. Residual symptoms after a treated major depressive disorder: in practice ambulatory observatory carried out of city. Encephale. 2003;5:438-444.
6. Katz MM, Tekell JL, Bowden CL, et al. Onset and early behavioral effects of pharmacologically different antidepressants and placebo in depression. Neuropsychopharmacology. 2004;29:566-579.

7. G. Parker,Australia

Gordon PARKER, MD, PhD, DSc, FRANZCP
Scientia Professor, University
of New South Wales
Executive Director, Black Dog Institute
Hospital Road
Prince of Wales Hospital
Randwick NSW 2031
AUSTRALIA
(e-mail: g.parker@unsw.edu.au)

It is generally suggested that antidepressant drugs “work” slowly, with Gershon1 noting the “established belief” that they take 2-6 weeks “to produce their antidepressant activity.” The devil almost certainly lies in the detail, as now detailed. First, the word “work” is capable of multiple definitions. In formal drug trials, “improvement” and “response” status are commonly operationalized as respective reductions in depression severity of at least 20% and 50%, while “remission” is viewed as the absence—or virtual absence—of any symptoms.

While remission—or a euthymic state—is the therapeutic goal, we are increasingly recognizing that full remission (the antidepressant drug has “worked”) is probably only achieved by a minority, with illustrative data from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study quantifying remission rates in the order of 30%, and a responder rate of 47% after up to 3 months of the antidepressant citalopram.2

Aggregating remitters, responders, and nonresponders in any sample of patients receiving antidepressant drugs therefore contributes to a fairly consistent trajectory pattern, and where, even after 2-3 months, group depression scores are still decreasing, so building the impression that antidepressants require an extended period before they “work.” However, such group trajectories are made up of formally defined “responders” and “nonresponders,” with the nonresponders distorting the quite differing trajectory of improvers and remitters.

If we then limit analysis to prediction of responder status, numerous studies3,4 have shown that for such responders, improvement status is achieved in the first week, with evidence of such early improvement also shown for those receiving electroconvulsive therapy.

Do such findings allow us to conclude that early improvement is to be expected of an effective antidepressant? No, possibly, and yes. Caveats to the expectation reflect two principal concerns. First, the majority of the studies examining improvement trajectories with antidepressant drugs have been weighted to patients with nonmelancholic disorders. For those with melancholic and psychotic depression, in the absence of clear data, the impression remains that any improvement may not be evident for several weeks (either as a consequence of the underlying pathogenesis or of the requirement for a certain drug dose to be achieved). Second, placebo responders tend to show evidence of “early onset” improvement, quantified by Quitkin and colleagues5 as occurring in the first 2 weeks, but generally associated with a subsequent relapse (unless spontaneous remission has been induced or promoted). Turning to the “yes” interpretation, the literature argues against the mythology that antidepressants require weeks or months to work. If a patient is likely to have their depression respond specifically to an antidepressant drug, then some indication of improvement should be evident in the first week or 2—and the improvement trajectory should bemaintained. If no such improvement, the clinician might well consider whether the drug dose is insufficient, if augmentation is required, or if another antidepressant drug or other therapeutic paradigm should be contemplated. While there is a minority of individuals who will show a delayed onset effect (ie, improvement occurring after weeks or even months), this appears relatively uncommon and should not dictate clinical practice. As noted previously,6 if late onset of improvement is a myth, “depressed patients need not necessarily be treated so patiently” by trialing an antidepressant drug for many months. _

References
1. Gershon S. Antidepressants: can we determine how quickly they work? Psychopharmacology Bulletin. 1995;31:21-22.
2. 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.
3. Katz MM, Kowlow SH, Maas JW, et al. The timing, specificity and clinical prediction of tricyclic drug effects in depression. Psychol Med. 1987;17:297-309.
4. Parker G, Blignault I. Psychosocial predictors of outcome in subjects with untreated depressive disorder. J Affect Disord. 1985;8:73-81.
5. Quitkin FM, Rabkin JG, Ross D, et al. Identification of true drug response to antidepressants: use of pattern analysis. Arch Gen Psychiatry. 1984;41:782-796.
6. Parker G. Recovery from depression: triggers and time patterns. ANZ J Psychiatry. 1996;30:442-444.

8. M. Roca,Spain

Miquel ROCA, MD
Professor of Psychiatry
University of Balearic Islands
Institut Universitari d’Investigació
en Ciències de la Salut (IUNICS)
Juan March Hospital
Palma de Mallorca, SPAIN
(e-mail: mroca@uib.es)

Clinical response and outcome in depression are related to multiple factors. Functional impairment and risk of suicide are two of the major problems at the beginning of treatment. The delayed antidepressant response theory has been dominant in biological and clinical research for more than three decades, despite the initial suggestion of Kuhn and the observation of the immediate inhibitory action of antidepressants on monoamine reuptake.

What is the definition of early improvement? A great number of methodological problems and limitations in the published papers on this topic are relevant: rating scales, frequency of assessments, statistical approaches… Most of the findings come from post-hoc analyses and meta-analyses of trials not specifically designed to detect the early onset of antidepressant action.1

Pooled analysis and systematic reviews or meta-analytic approaches give us a different perspective on the problem, and a positive answer to the controversial question. Forty-seven studies evaluating antidepressant drugs with established efficacy, performing weekly or biweekly evaluations, and presenting the time course of improvement, were included in a meta-analysis: 60% and 61% of the improvement that occurred on active medication or placebo, respectively, took place during the first 2 weeks of treatment. The results suggest that many patients demonstrate a true antidepressant response during the first or the second week of pharmacological treatment.2

Another systematic review andmeta-analysis of selective serotonin reuptake inhibitors (SSRIs) identified 50 randomized placebo-controlled trials of these drugs in the short-term treatment of unipolar depression in adults. The analysis supports the hypothesis that SSRIs begin to have observable beneficial effects during the first week of treatment. The effect was seen on the primary outcome of differences in depressive symptom rating scale scores and on the secondary outcome of achieving a 50% reduction in the score.3

A recent review using theMedline database (1966-2007) concluded that a certain group of experimental treatments can produce antidepressant response in a shorter period of time. The authors of the paper considered that a faster and sustained antidepressant response may prevent the neurobiological and psychosocial effects secondary to a recurrent or unremitting depressive episode, and could be a “new paradigm” in antidepressant treatment research.4

The Quitkin-Katz controversy5,6 has a recent chapter, with a post-hoc analysis of a placebo-controlled, randomized, double- blind study of patients with major depressive disorder treated for 8 weeks and then for another 6 months with duloxetine or escitalopram. Improvement at 2 weeks on the 17-item Hamilton Rating Scale for Depression (HAM-D17) significantly predicted remission. Early symptom changes were specific to treatment, with early response for the core depression factors of anxiety and motor activity for duloxetine, and anxiety for escitalopram. In conclusion, lack of early response on depression symptom subscales was highly predictive of a lack of sustained remission. The initial study by Katz et al6 did not include a placebo control group.

Overall, the findings of clinical studies and clinical experience have also confirmed the positive role of early response in predicting clinical outcome. Unfortunately, the published studies have not sufficiently assessed the behavioral changes that might accompany the early drug-induced changes in the monoamine systems. The real problem is how to relate early response to total response and to sustained remission; ie, depression as a chronic or long-term disease. Future research with findings from neuroimaging or pharmacogenetic studies in early response and remission is crucial to improve the outcome of affective patients and to give us a more appropriate answer. _

References
1. Leon AC, Blier P, Culpepper L, et al. An ideal trial to test differential onset of antidepressant effects. J Clin Psychiatry. 2001;62(suppl 4):34-36.
2. Posternak MA, Zimmerman M. Is there a delay in the antidepressant effect? A meta-analysis. J Clin Psychiatry. 2005;66:148-158.
3. Taylor MJ, Fremantle N, Geddes J, Zubin B. Early onset of selective serotonin reuptake inhibitor antidepressant action: systematic review and meta-analysis. Arch Gen Psychiatry. 2006;63:1217-1223.
4. Machado-Vieira R, Salvadore G, Luckenbaugh DA, Manji HK, Zarate CA. Rapid onset of antidepressant action: a new paradigm in the research and treatment of major depression. J Clin Psychiatry. 2008;69:946-958.
5. Quitkin FM, Rabkin JG, Ross D, Stewart JW. Identification of true drug response to antidepressants. Arch Gen Psychiatry. 1984;41:782-786.
6. Katz MM, Koslow SH, Maas JW, et al. The timing, specificity and clinical prediction of tricyclic drug effects in depression. Psychol Med. 1987;17:297-309.

9. S. Vahip,Turkey

Simavi VAHIP, MD
Professor of Psychiatry
Director, Affective Disorders Unit
Department of Psychiatry
Ege University Medicine Faculty
Bornova-Izmir 35100, TURKEY
(e-mail: simavi.vahip@ege.edu.tr)

Prediction of treatment response at an early stage may have many implications for patients and clinicians. Avoiding unnecessary exposure to ineffective drugs and lessening the negative consequences of depression are two of these. The question of the predictive value of early improvement or antidepressant response goes beyond an old discussion: the timing of the onset of antidepressant response. Three to four weeks’ delay was suggested as a common pattern for antidepressants in the early decades following their introduction. However, there are growing data to suggest early improvement—even as early as 1-2 weeks.1-3 There are two phenomena that have blurred our vision in this field: the placebo effect and the probability of spontaneous remission due to the episodic course of major depression. Differentiation of placebo effect from “true” drug response is the crucial discussion in this controversial issue.

About one third ofmajor depressive patients respond to placebo in drug trials. It is not possible to ignore this nonspecific effect for an active drug. It was suggested that the placebo effect was characterized by early onset of response and a fluctuating pattern, while true drug effect was characterized by a 2-week delay in onset and persistence of improvement, once achieved.4 Delayed persistent improvement was reported to occur about three times more commonly on drug than placebo in this study. The existence of delayed response and delayed persistent improvement has also been shown in another study.5

One method to validate “true” drug effect is to investigate the relationship between early and delayed response and longterm relapse, and several reports have provided some evidence for a delayed response of antidepressants as the “truedrug initial response pattern.”6 On the other hand, several investigators have provided considerable data for earlier “true” drug effect,1-3 and moreover, a possible predictive value of it for the treatment outcome.2,3

The main methodologies used to handle this controversial issue are open or randomized controlled trials using comparison groups. In recent times, advanced statistical approaches have been used for better evaluation, such as sensitivity, specificity, predictive values, area under the curve, and survival analysis. Two recent studies with these specific approaches shed light on the area. One of them analyzed data from a naturalistic study on a large sample of inpatients with major depression.3 Results supported early improvement in the first 2 weeks as predictor of later response and remission with high sensitivity in hospitalized patients. The second study2 was a meta-analysis carried out with 6562 patients. The authors concluded that early improvement with antidepressant medication can predict subsequent outcome with high sensitivity. Also, they stressed that there were high negative predictive values and little chance of stable response or remission in the absence of improvement within 2 weeks, and suggested that lack of improvement during the first 2 weeks of therapy might be considered as an indicator regarding earlier changes than conventionally thought.

A closer look at the literature indicates the heterogeneity of both “early/placebo response” and “delayed/true drug response.” Probably both include each other. There could be placebo responders among delayed responders and true drug responders among early responders. We cannot say that all questions are answered and all controversies clarified yet. We need to respect every single study and finding in the literature. Some controversies may be explained with different methodologies and different study populations. On the other hand, it is not possible to handle the issue without considering classification systems, possible subgroups in the category ofmajor depression, differential responses for subgroups, and the heterogeneity of individual responses to antidepressants. Studies designed to consider these issues and to target more representative populations in daily clinical practice may improve our knowledge in this crucial question. _

References
1. Taylor MJ, Freemantle N, Geddes JR, Bhagwagar Z. Early onset of selective serotonin reuptake inhibitor antidepressant action: systematic review and meta-analysis. Arch Gen Psychiatry. 2006;63:1217-1223.
2. Szegedi A, Jansen WT, van Willigenburg PP, van der Meulen E, Stassen HH, Thase ME. Early improvement in the first 2 weeks as a predictor of treatment outcome in patients with major depressive disorder: a meta-analysis including 6562 patients. J Clin Psychiatry. 2009;70:344-353.
3. Henkel V, Seemüller F, Obermeier M, et al. Does early improvement triggered by antidepressants predict response/remission? Analysis of data from a naturalistic study on a large sample of inpatients with major depression. J Affect Disord. 2009;115:439-449.
4. 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.
5. 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.
6. Nierenberg AA, Quitkin FM, Kremer C, Keller MB, Thase ME. Placebo-controlled continuation treatment with mirtazapine: acute pattern of response predict relapse. Neuropsychopharmacology. 2004;29:1012-1018.

10. M. Wong,Hong Kong

Michael WONG, MD
Chief of Service
Department of Psychiatry
Queen Mary Hospital
Room 210, Block J
102 Pok Fu Lam Road
HONG KONG
(e-mail: wongmcm@ha.org.hk)

This question will address the issue of the importance of feeling beneficial effects early in the treatment course for depression, the symptoms that through their improvement lead patients into remission, and the prediction of a response to treatment.

Patients with depression usually present to the doctor quite some time after they have begun to be depressed. There are different reasons for this delay in presentation and thus treatment. They may not think that they have an illness. Besides this, they may be so depressed that they do not have the motivation to seek medical treatment. On the other hand, they may be so pessimistic that they think they will not recover from the depressed mood. Quite often, they only come to the doctor when their symptoms are so severe that they have become very distressed or suicidal, or the significant people around them are also all feeling distressed. Therefore, it is important that treatment should aim to relieve their symptoms and the distress caused within a short period of time.

There are claims that certain antidepressant medications have a faster onset of action than others. However, the evidence is that all the commonly-used antidepressant medications take time to work, usually between 2 to 3 weeks. It is known that time is needed for the drug to reach a steady state in the blood. The serum level also has to be maintained so that changes in the neural circuitry beyond the monoamine receptors (eg, production of brain-derived neurotrophic factor, regeneration of neuronal networks) can occur, ultimately leading to elevation of the depressedmood. Unfortunately, side effects of the drugs will come up before the onset of the antidepressant action, making the patient more miserable and distressed.

Because of this time lag, the patientmay lose confidence in the treatment and even take it as a confirmation of their thoughts of hopelessness. Compliance with treatment, which is crucial, will be affected. Nevertheless, if some of the symptoms such as insomnia and agitation can be alleviated at the early phase of treatment, it certainly helps the patient to feel better and to motivate them for further treatment. The direction of further research in the development of antidepressant drugs should be in really achieving earlier onset of mood-elevating action. The future should probably include action beyond the monoamine receptor level or involve a mechanism of action on pathways other than those of the monoamines that we know presently.

Apart from drug treatment, certain psychosocial measures should also be implemented to help the patient to get better sooner. The depressed patient should be encouraged to engage in activities that they can cope with to build up their confidence and self-image. The level of activities can be increased gradually as the mood improves, in order to give the patient more positive experience. They should also undertake regular exercise, as there is evidence that exercise can stimulate the nerves in pathways related to mood regulation. All these measures will help the patient to move into remission sooner. _

11. R. M. Zaratiegui,Argentina

Rodolfo M. ZARATIEGUI, MD
Vice-Director, Postgraduate Institute
Argentinean Psychiatric Association (APSA)
Director, Psinapsys Centre
816-1/2, Street 48
B1900ANH, La Plata
ARGENTINA
(e-mail: rzaratie@gmail.com)

The present guidelines advise waiting between 4 to 6 weeks before changing or adapting treatment with antidepressants when there is not at least a partial response. We, as physicians, have got used to warning patients that they will have to wait several weeks to experience a significant improvement. Following research performed at Columbia University, it was considered for a long time that responses prior to a 3-week period of treatment were characteristic of the placebo effect and were not sustained over time (delayed- onset hypothesis).1 However, there have also been data that show that the antidepressant effect starts before this. The proportion of patients with a 50% drop in the score on a rating scale (usually Hamilton Rating Scale for Depression or the Montgomery–Åsberg Depression Rating Scale) is not the most suitable indicator of the beginning of antidepressant action. In general, there is agreement on the fact that improvement is already clinically noticeable with a 20%-25% drop.

In the last few years, four meta-analyses encompassing trials of several antidepressants versus placebo in more than 5000 patients have shown that active drugs have a higher percentage of responders even from the first week,2 and the same or a faster sustained response than placebo.3 Moreover, the major proportion of the difference seems to be during the first 2 weeks4 and is not due to the impact of the antidepressant’s sedative effect in the rating scale scores.

A helpful way to find out whether early improvement is a predictor of response or remission consists in calculation of its sensitivity, specificity, and predictive value. It was calculated in the most recent meta-analysis5 that a 20% improvement on the Hamilton Rating Scale for Depression in the second week was a sensitive indicator of a sustained response after the fourth week (81% to 87% sensitivity), both for antidepressants and for placebo, but not with very high specificity (50%). Quite notably, the absence of early improvement predicted lack of response in 90%, which would suggest that the second week should be a decisive moment at which to instigate any change in the treatment.

The contrast with the established notions seems to be due to the fact that the first studies were carried out with tricyclics and monoamine oxidase inhibitors, drugs of slower titration, and they were also statistically underpowered to detect differences in the first weeks. By contrast, in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, about a third of the responses happened after the sixth week. This is a group estimate without comparison with placebo, which does not mention when the improvement started but does remind us that it takes many weeks to be complete. However, in a naturalistic study in 795 inpatients, it was found that a 20% improvement within 2 weeks could predict 88% of the response at the end of the treatment, but with a low negative predictive value.6

Being able to earlier predict the effectiveness of an antidepressant implies earlier adjustment of the treatment, lower exposure to an inefficient drug, morbidity reduction, less workday loss, and lower family burden, as well as a better use of resources.

To conclude, we have reached a moment at which we should reconsider whether the delayed onset hypothesis must lead treatment in all cases. Under strict conditions, as in the efficacy studies, the evidence seems to justify a change or modification of the treatment toward the end of the second week if the patient does not show any improvement. However, we have to take into account that the effectiveness studies, such as STAR*D, remind us that those patients with comorbidities and other everyday practice–related characteristics tend to respond more slowly. _

References
1. Quitkin FM, Rabkin JG, Ross D, Stewart JW. Identification of true drug response to antidepressants. Use of pattern analysis. Arch Gen Psychiatry. 1984;41: 782-786.
2. Taylor MJ, Freemantle N, Geddes JR, Bhagwagar Z. Early onset of selective serotonin reuptake inhibitor antidepressant action: systematic review and metaanalysis. Arch Gen Psychiatry. 2006;63:1217-1223.
3. Papakostas GI, Perlis RH, ScaliaMJ, Petersen TJ, FavaM. Ameta-analysis of early sustained response rates between antidepressants and placebo for the treatment of major depressive disorder. J Clin Psychopharmacol. 2006;26:56-60.
4. Posternak MA, Zimmerman M. Is there a delay in the antidepressant effect? A meta-analysis. J Clin Psychiatry. 2005;66:148-158.
5. Szegedi A, Jansen WT, van Willigenburg AP, van der Meulen E, Stassen HH, Thase ME. Early improvement in the first 2 weeks as a predictor of treatment outcome in patients with major depressive disorder: a meta-analysis including 6562 patients. J Clin Psychiatry. 2009;70:344-353.
6. Henkel V, Seemüller F, Obermeier M, et al. Does early improvement triggered by antidepressants predict response/remission? Analysis of data from a naturalistic study on a large sample of inpatients with major depression. J Affect Disord. 2009;115:439-449.