Dosage Procedures for Adequate Trials of Antidepressants¹

Note: Full remission and recovery are the goals of adequate antidepressant trials. Response short of this should be considered unsatisfactory. Review diagnosis and psychosocial factors, and continue with the algorithm recommendations as indicated.²


Click here to go to other antidepressants: tricyclics, nefazodone, bupropion, venlafaxine, mirtazapine, MAOIs

Click here if this is a bipolar depression: dosage strategy will be somewhat different.

Begin with 20 mg citalopram, 20 mg fluoxetine, 50 mg sertraline, or 20 mg paroxetine. Use half this dose (or less) if the patient is elderly or considered likely to have side effects, although citalopram is usually started at 20 mg per day for most patients. Always be aware if the patient is on drugs metabolized by Cytochrome enzyme 2D6 (especially when using fluoxetine or paroxetine, and with perhaps lesser concern if the patient is going to receive citalopram or sertraline) or by IIIA3,4 (especially with fluoxetine and sertraline and with perhaps lesser concern if the patient is going to receive paroxetine, but there is no concern if using citalopram). Be prepared to start with a lower SSRI dose or make adjustments to the doses of the drugs that are substrates for the affected enzymes. Also use a lower dose if the patient is considered likely to have side effects, or if there are comorbid panic attacks associated with any anxiety disorder. To see a computer program that accesses a database of cytochrome enzyme interactions, covering most substrates, inhibitors, and inducers of the cytochromes, click here.

Does it help to routinely supplement with folate (folic acid, 0.5 mg per day) when you use an SSRI? Pretty good data says yes, if it's a female patient. Old studies showed that folate concentrations are reduced in many patients with major depression. Folate is a co-factor in biogenic amine biosynthesis. Coppen and Bailey, in a randomized, blinded, placebo controlled trial in 127 patients, found that the response rate (>50% drop in Hamilton) was 82% with folate plus with fluoxetine vs. 62% with placebo and fluoxetine.³ A later analysis of these data according to gender found that the difference was 94% vs. 61% in the women, while there was no difference in the men.(4) This impressive result needs large scale replication with more extensive measures of clinical and economic outcome. However, given that 0.5 mg of folate is the amount of folate in most multivitamin preparations, and since it is generally a good idea to take a multivitamin supplement anyway, it seems reasonable to have all female patients take a multivitamin that contains folate if they are not already doing so - when you start an SSRI.

If side effects develop as the patient begins with an SSRI, adjust the dose downward to find a tolerable dose, and then try to raise it later if possible. If there is rapid improvement followed by loss of benefit (full or partial) within two weeks, it might have been a placebo effect.(5) Or, less likely in such a short time, it might have been a true benefit that was lost. For either contingency, the dose may be increased by the amount of the original dose weekly up to the maximum recommended dose, if tolerated. [The maximum for routine use for citalopram is 60 mg, for fluoxetine it is 60 mg, with 80 mg for patients with comorbid OCD. The maximum for sertraline is 200 mg. The maximum for paroxetine is 50 mg.] The patient is then treated for another 3 to 5 weeks.

If there is absolutely no response after 4 weeks, there is unlikely to be subsequent improvement(6) and a dosage increase should occur, as above. If there is no response after two weeks, there is a moderate chance that there will be subsequent improvement(6) so generally it is best to wait.

In the inpatient setting, there will be intense pressure to make changes more rapidly, and it may not be possible to resist these pressures. However, shorter trials than the above (that do not lead to remission) can not be considered adequate for satisfying the conditions required for proceeding with the steps of the algorithms.

By "response" we mean a pattern of response that indicates true drug response rather than a non-specific, probably placebo initial response. Pattern analysis has suggested that early response (recovery in the first two weeks) and non-persistent response (response for a full week followed by an unimproved week) indicates probable placebo response. (7) If the placebo pattern occurs, remain suspicions that this is an unsatisfactory response. If there is a breakthrough of symptoms after an early response, or a pattern of non-persistent response occuring over several months of treatment, consider increasing the dose as recommended above for non-response or partial response. (8)

¹Osser DN, Patterson RD. Algorithms for the pharmacotherapy of depression: Part Two. Directions in Psychiatry 1998;18(4):319-334.

²Stahl SM. Why settle for silver, when you can go for gold? Response vs. recovery as the goal of antidepressant therapy. J Clin psychiatry 1999;60(4):213-4.

³Coppen A, Bailey J. Enhancement of the antidepressant action of fluoxetine by folic acid: a randomised, placebo controlled trial. J Affective Disorders 2000;60:121-130.

(4)Goodwin GM The addition of folic acid to fluoxetine for major depression increases response rates especially in women. Evidence-based mental health 2001;4:41

(5)Quitkin FM et al. Identification of a true drug response - use of pattern analysis. Arch Gen Psychiatry 1984; 41:782-786.

(6)Nierenberg AA et al. Early nonresponse to fluoxetine as a predictor of poor 8-week outcome. Am J Psychiatry 1995; 152(10):1500-1503.

(7)Stewart JW, Quitkin FM, McGrath PJ, et al. Use of pattern analysis to predict differential relapse of remitted patients with major depression during 1 year of treatment with fluoxetine or placebo. Arch Gen Psychiatry 1998;55:334-343.

(8)Byrne S, Rothschild AJ. Psychiatrists' responses to failure of maintenance therapy with antidepressants. Psychiatric Services 1997;48:835-837.

Dosage strategy when using SSRI's for bipolar depression.

Although essentially unstudied, a significant minority of "expert clinicians" advocate a "lower and slower" initial approach with pharmacotherapy in bipolar depression, although the majority favor using full doses as one would with non-bipolar depression. However, for a treatment-resistant bipolar depression, there was good agreement that the same strategy should be used for bipolar and non-bipolar depressions.¹ Patients with a clear history of antidepressant-induced cycling, or who are rapid cyclers (four or more episodes per year) without necessarily having a history of antidepressant-induced cycling, are perhaps the most appropriate patients to use the "lower and slower" approach.²

The majority of the expert clinicians would taper off the antidepressant more quickly in bipolar depression than in unipolar depression, to prevent cycling. Almost as many, however, would continue the antidepressant for the same length of time (6-12 months) before considering tapering it off.¹ Thus, the decision for any individual patient is a complicated one. Up to one third of subsequently treatment-refractory bipolar patients may be precipitated into mania by antidepressants: risk factors appear to be past history of antidepressant-induced mania, female gender, hypothyroidism, and bipolar II history.³ These risks may be significantly reduced by adding thyroid hormone, adding mood stabilizers, eliminating stimulants (including caffeine), alcohol, and steroids, and by eliminating the antidepressant as soon as possible.(4,5) The latter will be the last choice in patients with recurrent severe depressions or chronic dysphoria off antidepressants.(4)

It is important to discuss the risks benefits of the strategy chosen with the patient and document informed consent.

If you are treating a patient with bipolar depression, start a consultation and select bipolar depression as your diagnosis: much more detail will be provided.

Return to the adequate trial recommendations for non-bipolar.

¹Kahn DA, Carpenter D, Docherty JP, Frances A. The Expert Consensus Guideline Series: Treatment of bipolar disorder. J Clin Psychiatry 1996;57[suppl 12A]:53.

²Dantzler A, Osser DN. Algorithms for the pharmacotherapy of acute depression in patients with bipolar disorder. Psychiatric Annals 1999;29(5):270-284.

³Altshuler LL, Post RM, Leverich GS, et al. Antidepressant-induced mania and cycle acceleration: a controversy revisited. Am J Psychiatry 1995;152:1130-1138.

(4)Simpson HB, Horowitz GI, Liebowitz MR. General principles in the pharmacotherapy of antidepressant-induced rapid cycling: a case series. J Clin Psychopharmacol 1997;17:460-466.

(5)Boerlin HL, Gitlin MJ, Soellner LA, Hammen CL. Bipolar depression and antidepressant-induced mania: a naturalistic study. J Clin Psychiatry 1998;59:374-379.

Tricyclic Antidepressants

An adequate trial of a tricyclic antidepressant is defined by evidence of significant bioavailability in terms of plasma level and reasonable duration of the trial. The most recent meta-analysis of plasma level studies found that the therapeutic ranges are (shown below with the percentage response rate in range vs. out of range):¹

nortriptyline 58-148 ng/ml: 66% vs. 26% (window)

desipramine >115 ng/ml: 51% vs 15% (threshold)

imipramine (plus desipramine) 175-350 ng/ml: 67% vs. 39% (window)

amitriptyline (plus nortriptyline) 93-140 ng/ml: 50% vs. 30% (window)

all others: not well established

Plasma levels are to be used electively as one variable to consider in determining optimal dose. However, all patients should have at least one plasma level at some point in order to reduce the risk of the occult development of toxicity from high levels.² In the event of unsatisfactory response or excessive side effects, a plasma level becomes a primary criterion of an adequate trial for the purposes of satisfying conditions for proceeding to the next step of this algorithm. If another drug is added to the patient's tricyclic regimen that inhibits tricyclic metabolism, (such as an SSRI antidepressant or nefazodone), at least one follow-up plasma level should be obtained.

Clomipramine is also a tricyclic antidepressant, but adequacy of a trial of this medication is defined here according to its usage as a treatment for obsessive-compulsive disorder. In this literature, the typical dose is 250 mg per day. This (or the maximum dose tolerated) is the dosage criterion adopted in this algorithm for an adequate trial.

¹Perry PJ, Zeilmann C, Arndt S. Tricyclic antidepressant concentrations in plasma: an estimate of their sensitivity and specificity as a predictor of response. Journal of Clinical Psychopharmacology. 1994;14:230-240.

²Preskorn SH. Therapeutic drug monitoring with tricyclic antidepressants: a response. Journal of Clinical Psychopharmacology. 1994;14:277-278


Nefazodone is structurally related to trazodone and has some similarities and some differences with its cousin. However, there are occasional problems with liver failure with nefazodone. Initially, four patients were cited in two reports.¹ After 14 to 28 weeks of treatment, patients developed jaundice, followed by encephalopathy 2.5 to 6 weeks later, even after discontinuing nefazodone 2-7 days after symptoms began. Liver failure and death have been reported with other antidepressants including SSRIs, although no causal relationship was proven. A review of 99 case reports of hepatotoxicity from antidepressants concluded that the rate with nefazodone was considerably higher than other antidepressants, although at one in 250,000 to 300,000 nefazodone patients it was still extremely infrequent.² However, the FDA on Dec. 7, 2001 announced that there will be a black-box warning in the package insert of nefazodone, requiring that patients be warned of hepatotoxicity. Risk managers and experts seem to agree now that nefazodone should be used only after other medications have been tolerated poorly or been ineffective.³ Be sure to update whatever patient information materials you have been giving your patients when you prescribe nefazodone. Clinical features of hepatic injury can include anorexia, nausea, vomiting, abdominal pain, fatigue, dark uring, jaundice, and confusion, as well as more severe signs such as ascites and coma. If these occur, the patient should stop the medication and receive evaluation as soon as possible.³

Other properties of nefazodone include:

Until the black box warning came out, nefazodone used to be considered a first-line pharmacotherapy for depression based on its side effect profile. Many patients hate the sexual side effects of SSRI's, for example, and the early onset antianxiety effect and lack of need for a concomitant hypnotic (as is often the case with SSRI's) would be appreciated by many patients.

The median dosage associated with the best response to nefazodone is approximately 500 mg per day.(4) An adequate trial is defined as one in which this dose, or the maximum tolerated dose, is achieved and maintained for four to eight weeks. Begin at 50 mg twice a day for most patients, and increase by 50 to 100 mg every two days until reaching 300 to 400 mg. Hold there for two weeks, monitoring for side effects and therapeutic response. Further increases to 500 mg or if necessary 600 mg (the package insert maximum) occur if response is unsatisfactory and increased dose is tolerated. Then, another two to four weeks is allowed.

Some patients get more agitated, stimulated, or develop insomnia at the higher doses. This may be more likely if the patient is on another drug which is a cytochrome 2D6 inhibitor. Or they may have been on fluoxetine recently which (due to it's metabolite norfluoxetine) has a long half life and blocks 2D6 for many weeks. The reason may possibly have something to do with nefazodone's metabolite m-chloro-phenyl-piperidine (mCPP). This metabolite seems to induce anxiety, and is used in animal models and receptor studies of anxiety. This metabolite is metabolized by 2D6, so its levels may selectively increase if the patient is on a 2D6 inhibitor. The parent compound of nefazodone is primarily metabolized by 3A4.

Nefazodone is a potent inhibitor of cytochrome 3A4. Many drugs are substrates for this enzyme, so be aware of possible interactions. Check our Cytochrome Interaction Program, which enables you to enter the patient's drug regimen (or proposed drug regime) and obtain data on the possible interactions and their possible significance.

¹Aranda-Michel J, et al: Nefazodone-induced liver failure: report of three cases. Annals of Internal Medicine 1999;130:285-288

²Garcia-Pando AC et al. Hepatotoxicity associated with the new antidepressants. J Clin Psychiatry 2002 in press.

³Gelenberg AJ. Nefazodone hepatotoxicity: black box warning. Biological Therapies in Psychiatry 2002;25(1):2

(4)Janicak PJ, Preskorn SH, Davis JM, Ayd FJJ. Principles and Practice of Psychopharmacotherapy, 3d Edition. Williams and Wilkins, Chicago 2001


We recommend routine use of the sustained release preparation of this agent. Bupropion in all forms is contraindicated in patients with a history of an eating disorder.

The sustained release form of bupropion is usually begun with 150 mg once a day in the morning. It may be increased, if tolerated, in 4-7 days, to 150 mg bid (according to the package insert). The manufacturer's research¹ has made a convincing case that the seizure risk is significantly lower with the sustained release preparation, even though it was not studied at doses over 300 mg, where the risk might be higher. With bupropion SR, peak plasma concentrations are 50% of what they are with the same dose of the regular release form, with equivalent area under the plasma concentration curve for bupropion and its metabolites. The seizure rate for patients in combined acute and one year maintenance treatment was 0.15%. Patients with a history of an eating disorder or seizure disorder were excluded from the study. The package insert does state that the seizure rate on 400 mg of bupropion SR rises to 0.4%. This is not based on data from the sustained release trials but rather by extrapolation as a worst-case-scenario from experience with the regular release preparation.

If you use the regular-release preparation, therapeutic response to bupropion is generally seen at 300 to 450 mg (the package insert maximum) per day of the regular-release formulation, and dosage in this range is required for an adequate trial, unless side effects prevent an increase to this level. Begin with 75 mg in the morning, or 37.5 mg (by cutting unscored 75 mg tablet) in those with a high likelihood of sensitivity to side effects, and increase by 37.5 to 75 mg every other day until the recommended dosage is achieved or side effects intervene. Four to eight weeks is required to consider the trial adequate.

¹Dunner DL, Zisook S, Billow AA, et al. A prospective safety surveillance study for bupropion sustained-release in the treatment of depression. J Clin Psychiatry 1998;59:366-373.


As a norepinephrine and serotonin reuptake inhibitor, venlafaxine is a significant alternative to a tricyclic and to combined tricyclic/SSRI treatment. Its advantages would be the lack of tricyclic side effects such as anticholinergic effects, cardiac conduction slowing, alpha blockade, weight gain, and overdose danger. In melancholic, severely depressed patients, robust efficacy compared to placebo,¹ superior efficacy to SSRI's alone,² and superior efficacy to SSRIs in treatment-resistant depression³ are suggested in some published studies. Treatment-resistant patients may respond.(3a) Also, more rapid onset of clinical benefit, as has been suggested for combined tricylic/SSRI therapy, has been reported. (4) Although every new antidepressant claims to work more rapidly, the claim with venlafaxine seems to be based on more substantial evidence.

Clinicians' experience with venlafaxine was initially somewhat disappointing. Early side effects of nausea and agitation are particularly problematic, and seemed to occur at a much higher rate than suggested by the published studies. (Hence, some clinicians liked to refer to the drug as "side-Effexor.") The reason for these side effects may have to do with the fact that clinicans rarely used venlafaxine as first-line. It still is usually started after the patient has been on an SSRI. Since venlafaxine is metabolized by Cytochrome 2D6, and SSRI's inhibit metabolism of this enzyme, this could be the cause of prominent early side effects. Fluoxetine, with the long half life of its active metabolite norfluoxetine, might be particularly prone to still be present, even if the fluoxetine was stopped prior to starting the venlafaxine. There also might be pharmacodynamic synergism of the SSRI used previously and the venlafaxine. With the shorter acting SSRI's, if they are abuptly discontinued, there may be a discontinuation syndrome complicating interpretation of the apparent side effects reported by the patient after starting venlafaxine.

The side effects appear to be milder with the new sustained release form. More usage is encouraged in order to settle the question of whether this promising new drug deserves more widespread use for any situation where a tricyclic or tricyclic/SSRI combination seems indicated. A recent placebo-controlled study of first-line use in 384 patients found excellent results, and only 0.3% blood pressure elevation at doses below 200 mg per day.(6)

Beginning with a low dose of as little as 12.5 mg is suggested and slowly titrating up, as tolerated. The degree of response seems dose related, with the best results in treatment-resistant patients seen in doses from 200 to 375 mg per day.(4) Baseline blood pressure should be recorded and monitored periodically, as there is a small chance of hypertension expecially with doses over 300 mg. The presence of (treated or untreated) hypertension at baseline is not a predictor of who will get blood pressure elevation and is not a contraindication to using venlafaxine. Other side effects reported in studies are similar to those seen with SSRI's. Unlike the SSRI's, venlafaxine appears to have a small inhibitory effect on cytochrome enzymes. The new extended release version of venlafaxine is reported to induce less nausea and vomiting.(6)

As tolerated, titrate up to 150 mg per day over a two week period. If the response is unsatisfactory over the next four weeks, increase gradually to 250 mg and beyond to a maximum of 375 mg if necessary and tolerated. Baseline and periodic blood pressure measurements during titration are advised. Blood pressure elevations occur in a few percent of patients in doses over 200 mg and up to 10% of patients when doses go over 300 mg.

There is no increased risk of blood pressure elevation in patients with treated pre-existing hypertension.(7)

¹Guelfi JD et al. Effectiveness of venlafaxine in patients hospitalized for major depression and melancholia. J Clin Psychiatry 1995; 56:450-458.

²Clerc GE et al. A double-blind comparison of venlafaxine and fluoxetine in patients hospitalized for major depression and melancholia. Int Clin Psychopharmacology 1994; 9:139-143.

³Poirier MF, Boyer P. Venlafaxine and paroxetine in treatment-resistant depression. Double-blind, randomised comparison. Br J Psychiatry 1999;175(July):12-16.

(3a)Derivan et al. Venlafaxine: measuring the onset of antidepressant action. Psychopharmacology Bull 1995; 31(2):439-447.

(4)Nierenberg A et al. Venlafaxine for treatment-resistant unipolar depression. J Clin Psychopharmacol 1994; 14(6):419-423.

(5)Khan A, Upton V, Rudolph RL, et al. The use of venlafaxine in the treatment of major depression and major depression associated with anxiety: a dose-response study. J Clin Psychopharmacol 1998;18:19-25.

(6)Entsuah R. Benefit-risk analysis between venlafaxine XR and venlafaxine. . American Psychiatric Association Annual Meeting New Research. San Diego, California: American Psychiatric Association; 1997:187.

(7)Thase ME. Effects of venlafaxine on blood pressure: a meta-analysis of original data from 3744 depressed patients. J Clin Psychiatry 1998;59:502-508.


An adequate trial is defined as a dose of at least 30 mg per day for four to eight weeks, followed by another two to four weeks at 45 mg if the response is unsatisfactory. The usual starting dose is 15 mg.¹ At the higher doses, there seems paradoxically to be less weight gain and sedation than at 15 mg, perhaps due to the fact that antihistaminic activity diminishes in relation to dose.² Be sure to warn patients about the small risk of agranulocytosis, as the package insert demands. In post-marketing followup, the incidence has been extremely low.

¹Hopkins HS. Mirtazapine. Biological Therapies in Psychiatry Newsletter. 1997;20:2-4.

²Stahl S, Sussman N. Mirtazapine: a clinical profile. Primary Psychiatry. 1997;4:83.


An adequate trial is defined as at least 60 mg of phenelzine or 40 mg of tranylcypromine, or the maximum dose tolerated, for four to eight weeks. Isocarboxazid has become available again recently, as well. It is usually started at 10 mg twice a day and final dose ranges from 20 to 60 mg per day. Consult these references and the package inserts for full prescribing and dietary information.¹,²

¹Martin L, Bakish D, Joffe R. MAOI treatment of depression. In: Kennedy SH, ed. Clinical advances in monoamine oxidase inhibitor therapies. Washington, D.C.: American Psychiatric Press, Inc.; 1994:147-179.

²Gardner DM, Shulman KI, Walker SE, Tailor SAN. The making of a user friendly MAOI diet. Journal of Clinical Psychiatry. 1996;57:99-104.