You are leaving Medscape Education
Cancel Continue
Log in to save activities Your saved activities will show here so that you can easily access them whenever you're ready. Log in here CME & Education Log in to keep track of your credits.


Optimizing Efficacy and Tolerability of Antidepressant Therapy: Does Selectivity of Action Matter?

  • Authors: Chairperson: Michael E. Thase, MD; Faculty: Pedro L. Delgado, MD; Justine M. Kent, MD; Andrew A. Nierenberg, MD
Start Activity

Target Audience and Goal Statement

This program is intended for physicians and other healthcare professionals involved in the treatment of patients with depression and anxiety disorders.

Currently available antidepressants generally are viewed as comparably effective, despite widely disparate pharmacologic effects. Even among a class of agents, such as the selective serotonin reuptake inhibitor (SSRI), there is considerable pharmacologic and pharmacokinetic heterogeneity that may contribute to clinical differences in efficacy and tolerability. More recently, results from pooled and meta-analyses of completed data sets suggest that SSRIs may be somewhat less effective than multiple mechanism drugs, although prospective data are lacking.

This activity will describe the seven distinct mechanisms of antidepressant drugs currently in clinical use, assess the spectrum of efficacy of newer agents in depression and anxiety disorders, and evaluate the comparative effectiveness, safety and tolerability of various classes of drugs and individual agents.

Upon completion of this activity, participants will be able to:

  1. Discuss the distinct pharmacologic effects of different classes of antidepressants and how they contribute both to efficacy and side effects.
  2. Review the evidence pertaining to the comparative effectiveness of different antidepressants.
  3. Discuss the efficacy of SSRIs and other newer antidepressants for the treatment of a broad range of psychiatric disorders.
  4. Identify the risks and adverse events associated with newer antidepressants.
  5. Describe the unique pharmacologic properties of various antidepressants with individual patient characteristics to inform prescribing decisions.


  • Pedro L. Delgado, MD

    Department of Psychiatry, Case Western Reserve University, School of Medicine


    Disclosure: Grant support: AstraZeneca, Bristol-Myers Squibb, Eli Lilly and Co., Forest Pharmaceuticals, GlaxoSmithKline, Janssen Pharmaceutica, National Institute of Mental Health, Novartis, Organon, Inc., Pfizer, Inc., Pharmacia, Solvay Pharmaceuticals, Inc., and Wyeth;
    Advisory Board Member: Boehringer Ingelheim, Eli Lilly and Co., GlaxoSmithKline, Janssen Pharmaceutica, Organon, Inc., and Wyeth;
    Speakers' Bureau: Bristol-Myers Squibb, Eli Lilly and Co., GlaxoSmithKline, Janssen Pharmaceutica, Organon, Inc., Pharmacia, Solvay Pharmaceuticals, Inc., and Wyeth.

  • Justine M. Kent, MD

    Department of Psychiatry, Columbia University College of Physicians and Surgeons


    Disclosure: Consultant: Pfizer, Inc.

  • Andrew A. Nierenberg, MD

    Department of Psychiatry, Massachusetts General Hospital


    Disclosure: Grant support: Bristol-Myers Squibb, Cyberonics, Eli Lilly and Co., GlaxoSmithKline, Litchner, Pfizer Inc., and Wyeth;
    Consultant: Eli Lilly and Co., GlaxoSmithKline, Innapharma, Inc., Janssen Pharmaceutica, and Wyeth;
    Honoraria: Cyberonics, Eli Lilly and Co., GlaxoSmithKline, Pfizer, Inc., and Wyeth.

  • Michael E. Thase, MD

    Professor of Psychiatry, University of Pittsburgh School of Medicine; Chief, Division of Adult Academic Psychiatry, Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania.


    Disclosure: Grant Support: Cyberonics, Inc., Pharmacia, and Wyeth;
    Consultant: Bristol-Myers Squibb, Cephalon, Inc., Eli Lilly and Co., Forest Pharmaceuticals, GlaxoSmithKline, Novartis, Organon, Inc., Pfizer, Inc., Pharmacia, and Wyeth;
    Speakers' Bureau: Bristol-Myers Squibb, Eli Lilly and Co., Forest Pharmaceuticals, GlaxoSmithKline, Organon, Inc., Pfizer, Inc., Pharmacia, Solvay Pharmaceuticals, Inc., and Wyeth.

Accreditation Statements

    For Physicians

  • This activity has been planned and implemented in accordance with the Essential Areas and Policies of the Accreditation Council for Continuing Medical Education (ACCME). The Postgraduate Institute for Medicine is accredited by the Accreditation Council for Continuing Medical Education (ACCME) to provide continuing medical education for physicians.

    The Postgraduate Institute for Medicine designates this educational activity for a maximum of 3.0 category 1 credits toward the AMA Physician's Recognition Award. Each physician should claim only those credits that he/she actually spent in the activity.

    Contact This Provider

For questions regarding the content of this activity, contact the accredited provider for this CME/CE activity noted above. For technical assistance, contact [email protected]

Instructions for Participation and Credit

There are no fees for participating in or receiving credit for this online educational activity. For information on applicability and acceptance of continuing education credit for this activity, please consult your professional licensing board.

This activity is designed to be completed within the time designated on the title page; physicians should claim only those credits that reflect the time actually spent in the activity. To successfully earn credit, participants must complete the activity online during the valid credit period that is noted on the title page.

Follow these steps to earn CME/CE credit:

  1. Read the target audience, learning objectives, and author disclosures.
  2. Study the educational content online or printed out.
  3. Online, choose the best answer to each test question. To receive a certificate, you must receive a passing score as designated at the top of the test. Medscape encourages you to complete the Activity Evaluation to provide feedback for future programming.
You may now view or print the certificate from your CME/CE Tracker. You may print the certificate but you cannot alter it. Credits will be tallied in your CME/CE Tracker and archived for 5 years; at any point within this time period you can print out the tally as well as the certificates by accessing "Edit Your Profile" at the top of your Medscape homepage.

The credit that you receive is based on your user profile.


Optimizing Efficacy and Tolerability of Antidepressant Therapy: Does Selectivity of Action Matter?: Efficacy of Newer Antidepressants


Efficacy of Newer Antidepressants , Presented by Michael E. Thase, MD

Assessing Efficacy of Antidepressants: Clinician and FDA Perspectives

  • We're going to talk about differential efficacy, then spend some time to discuss why, for 40 years, we haven't had a clue empirically whether antidepressants really had differences in their clinical effectiveness or not. At the end, you'll become convinced about whether one antidepressant had more of an effect then others might actually have been true all along, and that the people who over-rely on evidence are, of course, overly influenced by artifacts and particular biases that can make evidence-based medicine go awry. Also maybe what, as a consumer with new research, you might begin to expect and demand in studies that can compare different antidepressants.

  • Efficacy of Newer Antidepressants

    Slide 1.

    Efficacy of Newer Antidepressants

    (Enlarge Slide)
  • When we look at our antidepressant drugs and we talk about what works, we essentially are all drawing from, first and foremost, our own clinical experiences. Our initial 10 prescriptive experiences with a new medicine greatly influence what we think about that medicine for the rest of time. If we simply use a new medicine with a handful of patients and nobody gets better we strongly believe that the medicine is not particularly useful.

    Keep that in mind because the vast majority of us prefer to use brand new medicines with people who have already done poorly, either from a tolerability standpoint or from an effectiveness standpoint, with established medications.

    So we are predisposed to see our new medicines as perhaps less effective than they might actually be simply based on our own selection bias, that we use newer medicines first to treat less treatable folks.

    We are influenced by marketing; you can see subtle increases in prescription rates following attendance at activities. The vast majority of us receive handouts and reprints from pharmaceutical reps and I can tell you with P = 1 in 1 million probability that the pharmaceutical rep is handing you a reprint that favors their product or presents their product, and they don't hand you reprints that dispute claims about the product, that's for the other pharmaceutical company reps to bring you. So we get marketed and counter-marketed almost every day in our practices, and there are advertisements in our journals.

    There is word-of-mouth; we each have colleagues, people who we trust and we sit next to at the dinner talks, and that is also an influential source of information. If the person tells you that something is great, you're more likely to try it; if something is terrible, you would be less likely to try it.

    You have lectures with continuing medical education attached to it as well as those on Thursday nights in a small restaurant in your hometown, which often don't have CME attached to it.

    And last in terms of how we prefer to learn, comes actually reading the fundamental evidence that is the basis of whether medicines are approved or not.

  • How Do Clinicians Assess Antidepressant Efficacy?

    Slide 2.

    How Do Clinicians Assess Antidepressant Efficacy?

    (Enlarge Slide)
  • What does it take to get approval? Fundamentally, it takes only 2 reasonably large studies with a placebo control in which the novel drug was definitely more effective than the placebo on an a priori dependent measure the Hamilton Depression Rating Scale. In order to get this approval companies have now realized that if the drug is reasonably effective, they need to do 8 studies in order to get their 2 positive studies so that they can bring their drug to market. At the end of the day, pharmaceutical companies are not interested, first and foremost, in science; they're interested in selling their product. They exist for their stockholders and for their employees to identify potentially commercially useful medications, bring them to market, and sell them.

    You probably didn't know that from Prozac (fluoxetine) to Lexapro (escitalopram) that if you look at all of the studies done for Food and Drug Administration (FDA) registration, that more than 50% of the studies failed to show that drug is more effective than placebo.

    So, about half of our studies failed to show that the drugs work and that's why we don't have reboxetine. They did 8 studies and they had the bad luck of only 1 of the 8 studies definitively showing that it works.

    Interestingly, there's a difference between a failed study and a negative study. A negative study tells you that the novel drug doesn't work but an established drug does work. The FDA will hold that against the manufacturers of the new medication; that's a worrisome thing because it doesn't happen that often. A failed study would be in which the drug doesn't work and the standard comparison drug doesn't work either.

    So there is a little tendency not to use standard drugs in their most vigorous or most aggressive titration schedule. If you don't know the new drug's best titration schedule you are putting it at a disadvantage against the known drug and if you push the known drug to its best point, you can't do that with confidence yet for the novel drug.

  • What Are the FDA Requirements for Establishing Antidepressant Efficacy?

    Slide 3.

    What Are the FDA Requirements for Establishing Antidepressant Efficacy?

    (Enlarge Slide)

Detecting Meaningful Differences in Antidepressants

  • How many of you were taught that if the antidepressant had passed FDA muster, it was more or less as effective as any other antidepressant; in essence, that there wasn't any 1 particular medication that stood out heads above the others? That is indeed what all of us have been taught, from an evidence-based standpoint, because that is what the evidence has suggested.

    This statement comes from he American Psychiatric Association's 2000 practice guideline and it's very definitive: approved medications all work and they all work more or less with the same average effect.

  • Are All Antidepressants Comparably Effective?

    Slide 4.

    Are All Antidepressants Comparably Effective?

    (Enlarge Slide)
  • No one has ever thought that the medicines work the same for each particular patient. There were individual differences and this has been believed going back to monoamine oxidase inhibitors (MAOIs) vs tricyclics. Nevertheless, what I'm going to tell you is that that statement is not true. Because we've only recently understood the different ways that our standard studies can go wrong and lead to something called a Type II error. A Type II error is when you declare that no difference exists when, there really was a difference and due to some aspects of study design, you weren't able to see the difference.

    The Type II error would be like you take a magnifying glass and look at a slide loaded with some horrible noxious bacteria and say there's nothing on the slide. Obviously the bacteria are there, you just can't see them with the magnifying glass. You need a higher power of illumination or magnification.

    In the contemporary studies using intent-to-treat model antidepressants, on average, have about 40% efficacy compared with 30% on placebo. So the average benefit of drug over placebo is only a relative proportion of about 33%. In another way, placebo effects account for up to 75% of drug efficacy. This is a good thing in the sense that depression is responsive to placebo expectancy factors.

    That's the basis that all therapeutic relationships start upon; it reflects our ability to educate, to instill hope, to support, and to provide a healing environment. But don't bad mouth the placebo effect; it is a very important part of medicine. It is just hard to find good medications when the placebo effect has grown from 20% to 40% in the most recent generation of studies. So the larger the placebo effect in a given study, the harder it is to show that a good drug works.

    Now what about that intent-to-treat model? Intention-to-treat means that you're counting everybody who drops out of our studies based on their last observation. Most people drop out because they're not getting better, or because they have an annoying side effect, or for other reasons unrelated to the treatment. So 8 or 9 out of 10 people drop out with high symptom scores. So intent-to-treat means that if you've got 30% dropouts, 27% of those patients are going to be listed as nonresponders in the trial and the average dropout percentage is in the low 30s.

    Our drugs only work 40% of the time in these new trials. The reason is because 30% of the evidence or so is stacked up against them, because of the intent-to-treat. That means even the best drug would not get more than 70% of the people better. Thirty percent would have had side effects or quit the study for some other reason.

    Another point is that when we measure treatment responses in studies, we actually measure the relief of symptoms and we measure the progressively smaller symptom burden. We do not actually measure wellness, sense of humor, energy in a positive sense, one's ability to concentrate in a positive sense; can you read for 30 minutes instead of only 15 minutes and so forth. There are some ceiling effects based on looking for the absence of illness as opposed to the presence of wellness.

    Getting back to that notion of what is the comparator drug doing in these studies, there are FDA registration reasons why the comparative drug often doesn't get its fairest shake but there are also commercial reasons. Remember, when we test for significance, one treatment vs another, we have something called the 2-tailed test. That means the difference could go either in favor of the new drug or in favor of the old drug. And if you have designed a powerful study to tell the difference between 2 good drugs, it is just as likely that that difference could be in favor of the old drug, as it is in favor of the new drug. Again, from the day your drug comes to the market, it does not help commerce to know that it's actually less effective than some standard comparison drugs.

    These studies are done large enough to find the difference between drug and placebo half the time. They are not done with a large enough groups of patients to find the difference between a good drug and a better drug. In fact, that difference won't be found anymore than 20% of the time the way we do our studies.

  • Why Is It Difficult to See Meaningful Differences in Controlled Trials of Antidepressants?

    Slide 5.

    Why Is It Difficult to See Meaningful Differences in Controlled Trials of Antidepressants?

    (Enlarge Slide)
  • Let me you show how hard it would be to find the difference between a good drug and a punitively better drug. Here's our 50% response rate; the lion's share of that response is attributable to nonspecific factors. There are dropouts that we've got to count in the intent-to-treat. Remember, somewhere between 1 and 2 out of 10 patients are not going to respond to any antidepressant that we pick for them so they would have no chance of responding to a potentially better medication. We have some patients who go on to develop treatment resistance. That means that if every patient who wasn't accounted for by these different categories would respond to the novel antidepressant, it would be at most only giving us somewhere between 12% and 20% more people better. And of course we have no wonder drugs; there would be no reason to think that every single other person would respond to that particular medication.

  • Distribution of Outcomes of Acute-Phase Treatment

    Slide 6.

    Distribution of Outcomes of Acute-Phase Treatment

    (Enlarge Slide)
  • In future if we really are serious about telling the difference between good and better medications then first of all, we have to do large sample studies. Three hundred is the bare minimum. We actually need to see studies comparing 600, 800, 900 patients per group. Our pharmaceutical friends will not do those studies; they don't have to. And our friends on the review committees haven't been particularly interested in funding those kinds of studies so we're kind of stuck. At this time we have no study large enough of our newer antidepressants to tell if one medication is really better than another.

    What can a poor researcher do in a sea of inadequate studies? We can combine the data from the studies. There are 2 ways to do that; one uses the study and the other uses the patients.

  • Research Strategies to Detect Meaningful Differences

    Slide 7.

    Research Strategies to Detect Meaningful Differences

    (Enlarge Slide)

Use of Meta-Analyses and Pooled Analyses to Assess Antidepressants

  • When you hear the term meta-analysis, if it's been used correctly, it could describe either of these approaches. But there is a more restrictive use and that is to describe the type of combination of study results in which the study's the unit of observation. So if there are 8 studies comparing treatment A and treatment B, you've got N equals 8 pairs of observation. This method works really well when you have a large and diverse group of studies.

    The alternative would be to take the data from every patient in every study of a coherent set of studies, and of course this would be preferred if there were only a small number of studies.

    Neither of these approaches is fully useful when data have been hidden, because a rule of nature is the more negative data there are on a particular treatment, the more studies for that treatment have been suppressed or hidden. So an attempt to pool results based only on published studies, for example, might accidentally overrate a treatment that is less effective simply because you didn't get your hands on all the data.

    Ultimately, it should not matter if you've combined 14 studies or if you do a meta-analysis of 14 studies; they should converge on the same point but there are relatively few comparisons in which we can use these 2 approaches simultaneously.

    Remember we got fluoxetine in 1987, and it took us 2 to 3 years to believe that it was the real thing; by 1990 it was our leading antidepressant and from 1991 and 1992, the other selective serotonin reuptake inhibitors (SSRIs) began to come in. And for 13 years now, they have been the leading antidepressants used in the United States and in most of the industrialized world as well.

  • Which Method Is Best?

    Slide 8.

    Which Method Is Best?

    (Enlarge Slide)
  • What do we know at the end of 16 years of experience with these medications? First, we cannot conclude that any one particular SSRI is more effective than the class itself. We do have evidence that the stereoisomer of citalopram may have an advantage over the racemic mixture, regular Celexa (citalopram), in a subgroup of patients with severe depression. But even that comparison doesn't hold across mild, moderate, and severe depression. I can tell you that they are comparable but they are not interchangeable and when a pharmacy benefit has required a patient to switch from one to the other for the capricious reasons of the benefit manager's. People can be doing well on one SSRI, be placed on another one, and plummet. And people can switch from one that is ineffective to ones that are effective. But at the end of the day, they did turn out to be comparably effective with the tricyclics, with the exceptions of number 1, the SSRIs work better for younger women, and number 2, the tricyclics work better for inpatients, at least when higher doses of certain tricyclics are used.

  • Meta-Analyses of SSRI Efficacy

    Slide 9.

    Meta-Analyses of SSRI Efficacy

    (Enlarge Slide)
  • This gets to the interchangeability question and if you allow for the fact that a lot of this work has been done under fairly shady circumstances, you can see that across the board people appear to have a good chance of responding to a second SSRI whether they didn't tolerate or didn't respond to the first. All but one of these studies are open label, but there's a double-blind one, in which you still got 52% comparing, in this case, mirtazapine against the second SSRI. So there is evidence that these medications can be useful in a second choice position and that they are not treating the exact same constellation of patients.

  • OneSSRI Failure Should Not Dictate a Class Switch

    Slide 10.

    One SSRI Failure Should Not Dictate a Class Switch

    (Enlarge Slide)
  • Here are the data from escitalopram vs racemic citalopram in a subgroup of patients; these are pooled from 3 studies. No single study would have contributed this exact answer, but when you start to work with groups of data, you can see differences. Here you see that in the more severely depressed patients that escitalopram carried a stronger therapeutic effect in 10-20-mg doses than citalopram in 20-40 mg doses. You might argue that that is because this is relatively more potent than 2 to 1, and that hypothesis remains to be proven. But here's an example within the class in which you're seeing a difference in a subgroup of patients.

  • Pooled Analysis: Efficacy Comparison of Escitalopram and Citalopram

    Slide 11.

    Pooled Analysis: Efficacy Comparison of Escitalopram and Citalopram

    (Enlarge Slide)
  • Here are the data on tricyclics vs SSRIs; here are the inpatient studies using amitriptyline and clomipramine and here are the inpatient studies that use nortriptyline, desipramine and maprotiline, the norepinephrine selective tricyclics. And interestingly, it suggests that it's not tricyclic-ness that caused them to win but the fact that they were dual reuptake inhibitors, serotonin as well as norepinephrine.

    Interestingly, if you compare across studies, you might also conclude that the dual reuptake tricyclics were more effective than the norepinephrine tricyclics.

    What does this mean, number needed to treat? This is an index of how practical this advantage would be in your day-to-day practice. This tells you that if this difference is applied to your everyday practice, for every 5 inpatients you would treat with a tertiary mean tricyclic as opposed to an SSRI, 1 more would get better. And that is a very large difference in the relative scheme of things. Here I did not apply an alternative metric called the number needed to harm, even when the tricyclics are more effective, would still favor the SSRIs. They are much, much more tolerable and safer than the tricyclics.

  • Advantage of "Dual Reuptake" TCAs Over SSRIs in Inpatient Studies

    Slide 12.

    Advantage of "Dual Reuptake" TCAs Over SSRIs in Inpatient Studies

    (Enlarge Slide)

Venlafaxine for Depression

  • What about newer dual reuptake inhibitors, medications that might deliver that benefit but at the same time have a safety profile more like the SSRIs?

  • Newer SNRIs

    Slide 13.

    Newer SNRIs

    (Enlarge Slide)
  • We have venlafaxine the first of the newer dual reuptake inhibitors, and here we see data from 29 studies, 19 of which looked at an SSRI. And if you've never seen this kind of meta-analysis, the vertical line here is the line of parity. So a finding in which the mean is right on the line would be a study in which there was essentially no difference. The more the mean moves to the right, it favors the comparison drug; the more it moves to the left, it favors in this case venlafaxine. Out of these 29 comparisons, looking at the SSRIs as a class, here we've got 13 fluoxetine studies, 1 Luvox (fluvoxamine) study, 4 Paxil (paroxetine) studies, and 1 Zoloft (sertraline) study.

    When you look at the class, you've got an advantage that you would see 1 time for about 16 patients that you would treat. This is such a large advantage that it carries with it the nonsignificant differences in the tricyclic studies and in the 3 studies of Remeron (mirtazapine) or Desyrel (trazodone). When you get to the end of the day here, you see an advantage that you would see about once in 16 new prescriptions in this particular approach to data.

  • Efficacy of Venlafaxine Compared With SSRIs and TCAs in Depression

    Slide 14.

    Efficacy of Venlafaxine Compared With SSRIs and TCAs in Depression

    (Enlarge Slide)
  • Here we've got my pooled analysis with the first 8 of those venlafaxine SSRI studies, and the number needed to treat them was about 10. So when you look at the individual patient data as opposed to the study results, you saw an even larger difference; may be that's an advantage when there is only a smaller amount of data. Even more importantly you were able to confirm that the SSRIs worked in these studies, because in many individual studies, the SSRI didn't work and in one of the individual studies, venlafaxine didn't work either. But here you see a nice stepwise advantage with reasonably large numbers; I think they are all a little more than 2000 patients.

    The newer findings are looking at 33 studies with the patients pooled. You'll notice that the placebo response remains the same, the SSRI comparison response remains the same, and the venlafaxine comparison remains significantly more effective although it has come down a little bit. That is because, whereas the majority of the patients in the first analysis received higher doses of the immediate relief formulation, the later analysis included some low-dose primary care studies as well as more of the studies using the extended-release form in lower doses.

    I do want to point out that even though we've got data from 33 studies, it is still heavily fluoxetine dependent, 19 of the 33 studies, and 27 of the 33 studies included either fluoxetine or paroxetine. So even though we've got this wealth of evidence, we're still quite thin in terms of some of the other widely used SSRIs.

    Notice that escitalopram was not in there, and that is because escitalopram was not available to be used as a comparison drug until this past year. You cannot compare somebody else's drug if it is not available for general use.

  • Pooled Analyses of Head-to-Head Double-Blind RCTs: Comparing Venlafaxine and SSRIs

    Slide 15.

    Pooled Analyses of Head-to-Head Double-Blind RCTs: Comparing Venlafaxine and SSRIs

    (Enlarge Slide)
  • Here are 2 studies done by Forest Laboratories using venlafaxine as the standard comparison drug. These are among the first ever to use a new dual reuptake inhibitor as the comparison and this study got a lot of press. It is a primary care study with relatively low doses, and whether you looked at observed cases or last observation carried forward, you had a dead heat. The dead heat between escitalopram and Effexor (venlafaxine), between escitalopram and venlafaxine, happened with an average dose of about 12 mg on escitalopram and I believe about 90 mg on venlafaxine. So this is a low-dose primary care study. If only we could be so lucky to treat such easy to treat patients.

  • Flexible-Dose Comparison of Escitalopram and Venlafaxine XR

    Slide 16.

    Flexible-Dose Comparison of Escitalopram and Venlafaxine XR

    (Enlarge Slide)
  • Now we have a full-dose psychiatry study in which patients were fixed at 225 mg/day or 20 mg/day, and we see again, no significant difference between escitalopram and venlafaxine. There is about a 4% trend favoring escitalopram although that is of course not anywhere near statistical significance in a study of 100 patients per group.

  • Fixed-Dose Comparison of Escitalopram and Venlafaxine XR

    Slide 17.

    Fixed-Dose Comparison of Escitalopram and Venlafaxine XR

    (Enlarge Slide)
  • There was one finding in this fixed dose study, 20 mg/day of escitalopram, 225 mg/day of venlafaxine extended release (XR), and there was a significant advantage in patients completing the trial. The dropout rate was substantially different here, about 13% on venlafaxine XR and about 3% onescitalopram. So that was an advantage that showed itself as a little, albeit a nonsignificant trend.

  • Tolerability Comparison of Escitalopram and Venlafaxine XR

    Slide 18.

    Tolerability Comparison of Escitalopram and Venlafaxine XR

    (Enlarge Slide)

Duloxetine for Depression

  • This is the newer dual reuptake inhibitor duloxetine. And here is a summary: 60 to 120 mg/day dosing (only at 60 mg twice a day dosing thereafter) and we have all 6 studies available that have been completed.

  • Duloxetine

    Slide 19.


    (Enlarge Slide)
  • The proof of concept would be that the dual reuptake inhibitor would beat the SSRIs, and there was a modest advantage, nowhere near statistically significant. Both drugs are effective, the number needed to treat on the left is 20; that's a relatively high number. But the duloxetine studies included a fairly wide range of patients so if we only looked at the upper half of the severity distribution then we see findings that mirror those in the first venlafaxine analysis; SSRIs beat placebo, as they should, and duloxetine beat the SSRIs with a number needed to treat of 9, almost identical to what we saw in the initial venlafaxine study. I'm not showing milnacipran data because there are too few studies published, the company has not made the data available, and the drug is not available anywhere in this hemisphere anyway.

  • Duloxetine Therapy Is Significantly More Effective Than SSRI Therapy

    Slide 20.

    Duloxetine Therapy Is Significantly More Effective Than SSRI Therapy

    (Enlarge Slide)

Bupropion for Depression

  • Let's look at 3 other new antidepressants.

  • Other Newer Antidepressants

    Slide 21.

    Other Newer Antidepressants

    (Enlarge Slide)
  • We've got bupropion, thought to be a norepinephrine and dopamine reuptake inhibitor. This medicine's upper dose is capped because doses above significantly lower seizure threshold. The company has shown that in addition to their 5 published studies with SSRIs they have done 7. Before we go to efficacy I want to say that this is the best antidepressant available in terms of not having sexual side effects, so it would have a distinguishing characteristic even if it had no greater efficacy.

  • Bupropion

    Slide 22.


    (Enlarge Slide)
  • And within a complete set of data looking at each individual patient, it is an absolute dead heat.

    And of great interest, these are newer studies and placebo is up to 35% remission. The advantage for bupropion vs placebo and the SSRIs vs placebo is identical to the advantage we saw for the SSRIs vs placebo in my first pooled analysis, in the second pooled analysis, and in the duloxetine pooled analysis. So now we are seeing 3 completely independent data sets showing identical results in terms of the size of the average difference between selective medications and placebo.

  • Bupropion and SSRIs Have Identical Efficacy

    Slide 23.

    Bupropion and SSRIs Have Identical Efficacy

    (Enlarge Slide)

Nefazodone, Mirtazapine for Depression

  • How about nefazodone? The drug is good for sleep, and it has relatively less sexual side effects than potent reuptake inhibitors. It's gotten bad press; this is the one most hurt by doctor selection bias. It is not widely used and as a result, the company has had no motivation to do the comparison analysis.

  • Nefazodone

    Slide 24.


    (Enlarge Slide)
  • Mirtazapine is also good for sleep; it is obviously a very potent antihistamine as well as a potent serotonin type 2 and 3 blocker. There are marked differences in tolerability favoring this drug or worse for this drug, depending if you're looking at older patients or younger patients.

  • Mirtazapine

    Slide 25.


    (Enlarge Slide)
  • Here are the results of the first 6 head-to-head trials. I'm showing a survival analysis here because at week 6 you have no difference in response rates between mirtazapine and the various SSRIs. But at every point between week 2 and 6, you do have significant differences.

    Whether this is due to that antihistaminic effect or whether it is due to the serotonin III, serotonin II effect, this medicine has a significant 1- to 2-week acceleration of benefit without having greater overall efficacy. So it stands uniquely to the others that I have talked about.

  • Mirtazapine vs SSRIs

    Slide 26.

    Mirtazapine vs SSRIs

    (Enlarge Slide)
  • Let me finish with the following summary:

    Number 1: there have always been meaningful differences between our antidepressants; I have used these same methods to demonstrate meaningful differences between the tricyclics and the MAOIs. Number 2: these meaningful relationships appear to be related to the mechanisms of action that these medications have. Mechanism of action is important because it conveys, first of all, a narrower or broader therapeutic effect but, on the other side of the ledger, either a narrower or broader profile of side effects. So a more broadly active medication is also going to carry with it a broader burden of tolerability.

    Generally, medications that are simultaneously effective against norepinephrine and serotonin systems show an advantage in treating severe depression, and they may be more effective in ambulatory studies if and when they are comparably tolerable and that's where the newest venlafaxine/escitalopram study really stands out.

    Finally, give up your therapeutic nihilism if you would like to be optimistic for a change; we think not only about how medications work but the probability that the patients will get better. This gives us more optimism in terms of developing even more effective medications in the future.

  • Summary

    Slide 27.


    (Enlarge Slide)