Skip Navigation
Department of Health and Human Services

Return to Web Conference

What works for depression?
An Evidence-based Comparison of Treatment Options


Slide 1

CHRISTEN HORN: Well thank you all for joining us. My name is Christen, and I will be the moderator for today's Web conference.

Slide 2

Here’s an overview of what will be covered during today's Web conference. We'll begin with a very brief introduction to the Agency for Healthcare Research and Quality, or AHRQ, its Effective Health Care Program, and comparative effectiveness research from Bruce Seeman from the Office of Communications and Knowledge Transfer at AHRQ. Then Dr. Gartlehner will provide an update on second-generation antidepressants for treating adult depression.

Dr. Gartlehner is the primary author of the research summary titled Second-Generation Antidepressants and the Pharmacologic Treatment of Adult Depression, an update of the 2007 comparative effectiveness review, making him particularly qualified to speak on this topic. He is also the associate director of the AHRQ-funded RTI, or Research Triangle Institute, at the University of North Carolina Evidence-Based Practice Center.

Slide 3

Before we begin, please note that your audio line has been muted to prevent any background noise during the presentation. You may ask a question to all panelists and to the presenter at any time, but please note that all questions will be answered midway through and at the very end of the presentation.

If you're experiencing technical issues, as I mentioned earlier, you also may use the WebEx Q&A function to request help. And just so you know, on the screen there is a box with a question mark and that is the Q&A button that you can use.

Lastly, we would appreciate your feedback on this session. Share your thoughts using a short evaluation form that will appear on your screen once the meeting has ended. Just look for the tab marked evaluation on the pop-up message. We promise there are very few questions and will only take a couple of minutes to complete.

And with that, I will pass it on to Bruce.

Slide 4

Here you go. After you Bruce.

BRUCE SEEMAN: Thank you, Christen, I appreciate it. Welcome, everybody.

As Christen mentioned I'm Bruce Seeman from the communications office at the Agency for Healthcare Research and Quality. I know we're all eager to hear from Dr. Gartlehner, but I thought I would just offer a little bit of context about AHRQ and the origins of this research.

So the Agency for Healthcare Research and Quality's mission is to improve the quality, safety, efficiency, and effectiveness of health care for all Americans. Our research is, 80% of it is of AHRQ's budget is invested in grants and contracts focused on improving health care.

An important part of AHRQ’s work comes out of the Effective Health Care Program, which was created in 2003 under the Medicare Modernization Act. The EHC Program provides current unbiased evidence on clinical effectiveness of health care interventions. It focuses on patient-centered outcomes. It helps consumers, providers, and policy makers to make informed choices. It does not make treatment recommendations, and the long-term goal of the EHC Program is to improve health care quality and patient health outcomes through informed decisionmaking by patients, providers, and policy makers.

Slide 5

Over time, AHRQ has established the Effective Health Care Program as a leader in the field of comparative effectiveness research. Just a few words about how we define CER here at AHRQ.

Comparative effectiveness research is a type of patient-centered outcomes research, which compares drugs, medical devices, tests, surgeries, or ways to deliver health care, so that patients and their families can make more informed choices. The findings are descriptive, not prescriptive, and are not intended as tools, or I'm sorry, and are intended as tools for informed decisionmaking, not recommendations. And finally, findings highlight current evidence about effectiveness, risks, and side effects.

So those are a few words about AHRQ and comparative effectiveness research, and with that, I will turn it over to Dr. Gartlehner.

GERALD GARTLEHNER: Okay, thank you very much, Bruce.

As Christen already said, my name is Gerald Gartlehner. I'm the associate director of the RTI University of North Carolina Evidence-Based Practice Center,...

Slide 6

and I would like to...

Slide 7

start my presentation with a quick slide on funding support and a disclaimer. The report was funded by the Agency for Healthcare Research and Quality, and findings and conclusions that I'm about to present are those of the authors of the report.

The information is intended to help with decisionmaking but should not replace clinical judgment in any way.

Slide 8

So in the next 30, 35 minutes, I would like to first discuss some background about second-generation antidepressants and about the burden of disease of depressive disorders.

Then, I will summarize comparative data on the efficacy, effectiveness, and harms of 13 new antidepressants. And specifically, I will address the overall comparative effectiveness of treatments during acute phase treatment, whether there are any differences in effectiveness among second-generation antidepressants in patients with unresponsive or recurrent disease, whether there are any differences in effectiveness in patients with depression and accompanying symptoms.

I will discuss the comparative risk for adverse effects for harms, and discuss effectiveness and adverse events in different patient subpopulations. And then at the end, I will try to point out some clinical implications and draw some conclusions.

Slide 9

Depressive disorders, such as major depressive disorder, dysthymia, or subsyndromal depression, are serious, disabling illnesses causing a vast burden of disease. For example, major depressive disorder affects about 1 in 5 adults at some point during their lifetimes. And in any given year, about 7% of the U.S. population suffers from a depressive episode. And already in 2000, the economic burden of depressive disorders in the United States was estimated to be around 83 billion dollars. And more than 30% of these costs actually came from direct medical expenses.

Women are about twice as likely as men to develop a depressive disorder. And second-generation antidepressants, such as SSRI, select serotonin reuptake inhibitors, or norepinephrine reuptake inhibitors, dominate the medical management of depressive disorders, and they have mostly replaced first-generation antidepressants, such as tricyclic antidepressants.

Slide 10

In 2011, antidepressants have become the most commonly prescribed medications in the United States, with more than 260 million dispensed prescriptions, and this makes antidepressants ranked seventh with respect to spending with about 11 billion dollars spent in 2011.

For individual patients, according to the U.S. Consumers Union, monthly costs of antidepressants range from 25 dollars to up to 700 dollars, depending on the chosen medication.

Slide 11

And because antidepressants are very commonly prescribed drugs with a lot of revenue involved, second-generation antidepressants are also one of the drug classes with a high number of so-called “me-too” drugs.

So, “me-too” drugs are structurally very similar to already-known drugs with only minor chemical differences. These “me-too” drugs are no therapeutic innovations despite what marketing often says. The primary goal of “me-too” drugs really is to gain market shares for the producer of the drug.

Slide 12

And this situation leads to a mind-boggling network of studies comparing second-generation antidepressants with one another. It is almost impossible for an individual person to really make any sense out of this confusing network of studies, and really to be able to figure out whether one specific drug is really more efficacious than the others.

Slide 13

And consequently, over the past years there has been a lot of controversy about differences among second-generation antidepressants. About six years ago in 2011, I’m sorry, about in 2007, we published a study that found no substantial differences in efficacy among second-generation antidepressants.

A few months later, a British-Italian group published a very similar study in The Lancet that pretty much stated the opposite of what we were saying, namely that escitalopram and sertraline, have a more beneficial efficacy harms profile than other antidepressants. So both studies were looking at the same evidence. This rendered a lot of letters to the editor and commentaries.

Ultimately we thought that we had a more rigorous methodological approach than the other group, but what we did then for the

Slide 14

report that I am about to present, is that we still used our more rigorous methodological approach, but we employed the same statistical model as the researchers of The Lancet paper, because this was the main difference between the two studies and we just wanted to see whether we could reproduce their findings and see whether escitalopram and sertraline are really more efficacious and, in a nutshell, we could not reproduce their findings.

Slide 15

So in this report, our population of interest were adult outpatients with major depressive disorder, with dysthymia, or subsubsyndromal depression as defined by the DSM-IV.

Slide 16

And for major depressive disorder we assessed the comparative effectiveness of antidepressants for all phases of the disease as presented here on this slide. We were interested in the acute phase, so the acute phase is the phase where patients present, usually presents to the clinician with symptoms.

It's up to 12 weeks. The treatment goal of the acute phase is response or remission, then we also assess the continuation phase from 4 to 9 months during which relapse can occur, and also the maintenance phase, starting at 10 months.

The goal of maintenance treatment is to decrease the likelihood of a subsequent depressive episode.

Slide 17

We included 13 second-generation antidepressants. These are select serotonin reuptake inhibitors, or SSRIs, serotonin norepinephrine reuptake inhibitors, or SNRIs, but also other drugs with related mechanisms of action that target neurotransmitters.

At the time when our report was finished, the latest antidepressant, vilazodone pristiq was not approved yet, and therefore it is also not included in our report. Many of these drugs are now available generically.

Although the newer agents, such as desvenlafaxine and duloxetine, still have remaining patent protection. Except for fluvoxamine, all second-generation antidepressants are approved for the treatment of major depressive disorder. Fluvoxamine only has been approved for the treatment of obsessive compulsive disorder.

Slide 18

I don't want to talk too much about methods. We followed standard international guidance for comparative effectiveness reviews. We searched multiple electronic databases and also tried to find unpublished studies.

Because not all antidepressants have been compared in studies with one another, we used advanced statistical techniques, such as network meta-analyses to estimate comparisons of drugs that have never been compared directly with each other.

Slide 19

For our results, we used a 4-part hierarchy to rate the strength of the evidence. And what strength of evidence basically means is, it reflects the confidence that we have in the findings.

So for example, high strength of evidence means that we have high confidence that the evidence actually reflects the true effect. And that further research is very unlikely to change the estimate of these, of this effect.

On the other end of the spectrum is low, where we think that further research is very likely to change the estimate of the effect, and insufficient means that evidence either is unavailable or does not permit a conclusion.

And I am mentioning this classification because when I'm presenting results, I will also talk about the strength of evidence for each finding.

Slide 20

So in our report we were interested in both beneficial and harmful outcomes. Outcomes of interest for efficacy and effectiveness for example were response, remission, speed of response, relapse, quality of life, and in general, any health outcome that is patient relevant.

For safety and tolerability, we included overall adverse events, withdrawals due to adverse events, serious adverse events, and a number of specific adverse events.

Slide 21

So overall, we reviewed more than 6,000 abstracts and included 228 studies. These were randomized control trials but also observational and non-randomized studies.

Slide 22

So what are the findings? I'm focusing on major depressive disorder first.

In general, the evidence showed that more than half of patients with major depressive disorder do not fully recover during 12 weeks of treatment.

37% did not even respond to the treatment during 6 to 12 weeks, and 53% did not achieve remission. And the problem with this really is that currently the evidence is insufficient to determine factors that can reliably predict response or non-response in individual patients.

Slide 23

Overall, however, our results indicate that second-generation antidepressants have very similar efficacy and effectiveness. Some of our meta-analyses showed very small statistically significant differences between some drugs. Clinically, however, we believe that these differences are likely not relevant. The overall strength of evidence supporting no substantial differences in efficacy and effectiveness was graded as moderate.

Quality of life functional capacity was infrequently assessed, usually only a secondary outcome. The available studies, however, indicated no differences with respect to health-related quality of life either, and the strength of evidence is also moderate.

Seven studies reported that Mirtazapine has a significantly faster onset of action during the first two weeks of treatment than some SSRIs. All of these studies, however, were funded by the maker of Mirtazapine so has to be viewed cautiously.

After four weeks of treatment, response rates then were similar again compared with other SSRIs, and the strength of evidence for this finding is also moderate.

Slide 24

Also, we did not find any differences in efficacy and effectiveness between immediate and extended release medications. For example, between Fluoxetine daily and fluoxetine weekly or paroxetine immediate release or Paroxetine controlled release. The evidence with respect to better adherence and persistence of patients using extended release medications was mixed. Some studies showed better adherence, others did not.

Slide 25

So in patients who have achieved remission during the acute phase treatment, so these are fewer than 50%, the treatment goal then becomes the prevention of relapse and the maintenance of the remission, and the American College of Physicians, for example, recommends that clinicians continue treatment for 4 to 9 months after a satisfactory response in patients with the first episode of major depressive disorder.

And just as for acute phase treatment, we found moderate strength evidence that most second-generation antidepressants effectively maintain remission with similar efficacy.

Slide 26

If patients fail to recover with the initial antidepressant treatment, so this is about 1 in 3 patients, then switching to a different antidepressant becomes an option. And when switching to a second-line antidepressant, there is low strength evidence that venlafaxine may be modestly superior to SSRIs. The strength of evidence is low indicating the high uncertainty that really comes with this finding.

Other strategies that we did not consider in this report are augmenting the treatment with a second medication, or adding an evidence-based psychotherapy to the depressant, antidepressant regimen.

Slide 27

So I would like to pause now briefly to see whether there have been any questions so far. Christen, has there been?

MS. HORN: Actually, we have a couple. The first question is, “So this study looked exclusively at second-generation medications.

Why did you omit first-generation medications and what are the benefits of second-generation compared to first?”

DR. GARTLEHNER: Okay. Yeah, the study exclusively looked at second-generation antidepressants. AHRQ has commissioned a report comparing first- and second-generation antidepressants in the early 2000s, I believe. This report showed that the efficacy is very similar between the first- and second-generation antidepressants. The side effect profile, however, is much more in favor of second-generation antidepressants.

The problem with the first-generation antidepressants is that the therapeutic dose is very close to the toxic dose, and particularly with respect to toxicity when taken in overdose second-generation are just much safer than first-generation antidepressants. So this also is the reason why second-generation antidepressants really have become the treatment option of choice.

MS. HORN: Great. And Mark asks, “Did you include seasonal effective disorder?”

DR. GARTLEHNER: We did not include seasonal effective disorder.

MS. HORN: Okay.

DR. GARTLEHNER: But, there is a Cochran report on the treatment of seasonal effective disorder with second-generation antidepressants.

MS. HORN: Great. Thank you. And one last question, unless there are others, again reminder please use your Q&A function.

“The research summary states that a limitation of your conclusions was that they have been derived primarily from efficacy trials. Are there plans for more effectiveness trials to be conducted for comparison purposes?”

DR. GARTLEHNER: Okay. Yeah. Most of the studies that we included, most of studies that we found were so-called efficacy studies. So these are studies with highly selected populations in very controlled environments, and these efficacy studies always have limited applicability to the real world.

We found a few effectiveness studies, some from the United States, others from Europe, and the reassuring part was that effectiveness and efficacy studies actually showed the same treatment effects and also the same magnitude of treatment effects. I am not aware of any large effectiveness studies that are currently being conducted.

MS. HORN: Okay. Great. I think we can move on.


MS. HORN: I'm sorry. There's another question that came in.
Lawrence asks “Because the FDA allows [INAUDIBLE] for the active ingredient in a medication this creates problems in efficacy with brand versus generic.”

So more of a statement, do you want to respond?

DR. GARTLEHNER: Well, I think Mark is correct. Bio availability of generics might vary a bit, and if clinicians switch to generics probably patients should be monitored to really make sure that the antidepressants work or that maybe the dosage needs to be increased.

MS. HORN: Thank you very much. I think we can continue on.


Slide 28

So the next portion of my presentation addresses the comparative benefits of medications for patients with depression and accompanying symptoms. We found a number of accompanying symptoms that have been addressed in the literature, such as anxiety, insomnia, low energy, melancholia, pain, psychomotor change, and somatization.

This set does not represent a complete list of symptoms that commonly accompany depression. For example, we did not find any studies addressing appetite change for example, a common symptom that accompanies depressed patients.

Slide 29

So, we looked at the evidence from two different perspectives.

First, are there any differences among second-generation antidepressants for the treatment of depression in patients with accompanying symptoms? And second, are there any differences among second-generation antidepressants for the treatment of accompanying symptoms in patients with depression? And with respect to treating depression, we found similar efficacy amongst antidepressants.

However, as you can see on the slide for some symptom clusters of the evidence of really insufficient to draw conclusions.

Slide 30

With respect to the treatment of accompanying symptoms in depressed patients, we also did not find any substantial differences in efficacy, and again for many symptoms the strength of evidence was insufficient. For the treatment of accompanying anxiety and pain the strength of evidence was moderate that the efficacy is similar.

And this is remarkable because Fluoxetine, for example, has been heavily marketed with for the treatment of depression with accompanying pain. You might have seen the “Depression Hurts” campaign, but in our analysis and we re-ran meta-analyses, we did not find any differences in efficacy for Fluoxetine compared with other second-generation antidepressants with respect to the treatment of pain in depressed patients.

Slide 31

So, the next portion of the presentation then addresses whether second-generation antidepressants differ in safety or adverse effects. Data here included outcomes from head-to-head trials from placebo-controlled trials but also from observational studies. And we focused on three main outcomes: the overall comparative risk or specific adverse event of specific adverse events; discontinuation because of adverse events; and then the risk of severe adverse events.

Slide 32

So in the studies that we reviewed, about 63% of patients in the efficacy trials experienced at least one adverse event. And common adverse events were constipation, diarrhea, dizziness, headache, nausea, sexual dysfunction, and, and somnolence. Nausea and vomiting were actually the most common reasons for discontinuation in the efficacy studies.

Slide 33

Overall, there is higher strength evidence that the overall rates of adverse events are similar among second-generation antidepressants, although the incidence of specific adverse events differed across antidepressants, and I will address these differences in, in the following slides.

Slide 34

So we found high strength evidence that venlafaxine has a substantially higher risk of nausea and vomiting than SSRIs, low strength evidence as such paroxetine immediate release may lead to higher rates of nausea than paroxetine controlled release. The strength was high that mirtazapine is associated with more weight gain than SSRIs, citalopram was associated with higher rates of diarrhea than comparative drugs.

Slide 35

The strength of evidence was moderate that trazodone is associated with a higher incidence of somnolence. High strength evidence says that bupropion has fewer sexual side effects than SSRIs. And moderate strength evidence indicates that paroxetine had the highest rate of sexual side effects when compared with other SSRIs, and there's also low strength evidence that indicates that sexual side effects may occur more frequently in men than in women.

Slide 36

When compared with SSRIs, the highest discontinuation rates because of adverse, because of adverse events were seen with duloxetine and venlafaxine.

Venlafaxine, on the other hand, had the lowest discontinuation rates due to lack of efficacy. The highest rates of withdrawal symptoms were reported after discontinuation of paroxetine and venlafaxine. And fluoxetine had the lowest rate of withdrawal symptoms.

Slide 37

We also looked at rare but serious adverse events, and rare but serious adverse events such as suicidality, so suicidal thinking and suicidal behavior, seizures, cardiovascular events, serotonin syndrome. They really are a concern for all second-generation antidepressants.

The existing evidence, however, is currently insufficient to really compare the risk of these rare but severe adverse events. And particularly suicidal thinking and suicides have been discussed a lot in the media.

The evidence based currently suggests that only patients between 15 and 24 years are at increased risk of suicidality, by contrast patients who are 55 years or older appear to have a decrease in suicidality after antidepressant treatment.

Slide 38

With respect to differences in subgroups, the majority of trials found no differences in efficacy in elderly patients compared with younger patients. There might be some differences in the risk for adverse, in the risk for adverse events. Elderly patients are more susceptible to experience adverse events than younger patients.

Slide 39

The evidence about differences in efficacy effectiveness or risk of harms for subgroups with respect to sex, race, ethnicities, and common co-morbidities really was insufficient, we couldn't draw any conclusions about that.

Slide 40

As mentioned in the beginning we were also interested in patients with dysthymia and subsyndromal depression. However, no double-blind RCTs are available currently to compare second-generation antidepressants with one another.

And also the evidence base on the general efficacy of second-generation antidepressants for the treatment of dysthymia and subsyndromal depression really is cause, it is not well established whether these drugs work at all for treating dysthymia and subsyndromal depression. Findings of the studies that we had really were mixed.

Slide 41

So what does this all mean? What are the clinical implications of our findings? Well, for treating major depressive disorder, the evidence does not warrant the choice of one second-generation antidepressant over another based on differences in efficacy, effectiveness, or accompanying symptoms.

And although second-generation antidepressants are similar in efficacy, they cannot be considered identical drugs. There are differences, and these differences primarily are with respect to onset of action and adverse events, and these are differences that probably need to be taken into consideration for choosing a medication.

Slide 42

So differences in adverse events primarily but also costs might influence the choice of a medication for an individual patient.

Clinicians need to explore patient preferences about dosing regiments, about the level of acceptance that individual patients have for various adverse events, and they maybe also consider costs when choosing a treatment.

And given the difficulty in predicting what medication will be both efficacious and tolerated by an individual patient and given that about 50% will not achieve remission during acute phase treatment, it is probably prudent for clinicians to be familiar with a broad spectrum of antidepressants.

And so this slide ends my presentation. Thank you for your attention, and I think we are open for questions again.

MS. HORN: That's right. Thank you so much.

And we have a few. The first is from Avril, “Is there a drug of choice for PMDD, and if there is, has there been any trials comparing its effectiveness against some of the natural remedies, such as chasteberry, [primrose oil], and the like?”

DR. GARTLEHNER: Okay. In, in our report, we did not review premenstrual dysphoric disorder. We did a report for the Oregon Drug Effectiveness Review Project a few years ago.

There are very few studies that assess premenstrual dysphoric disorder and antidepressants. I have to say I'm not aware of any study or any systematic reviews that looked into natural remedies, such as, as you mentioned, chasteberry or primrose oil.

MS. HORN: Great. And a question from Mark, “Did you find evidence that second-generation antidepressants are or are not helpful adjuncts for pain management?”

DR. GARTLEHNER: Well, they all have some; they all lead to some pain relief.

What we found simply was that there is no, that there are no substantial differences in the amount of pain relief that you can achieve with second-generation antidepressants, so primarily we could not reproduce that veloxatine is more efficacious than other second-generation antidepressants when it comes to the treatment of accompanying pain.

MS. HORN: Okay. Next question.

“What should doctors or other health professionals take into consideration when prescribing antidepressant medications to patients given the findings of this research?” Okay.

DR. GARTLEHNER: Yeah, so given that the efficacy and effectiveness is similar across these 13 second-generation antidepressants, I think clinicians really can focus on adverse events, and the acceptability that patients have toward specific adverse events, so younger patients for example might be more concerned about sexual dysfunction or about weight gain.

Older patients, for older patients it might be more problematic if they have a high risk for diarrhea or vomiting, so these are things that need to be taken into consideration.

Also the dosing regimen probably should be taken into consideration. Some second-generation antidepressants like bupropion has to be taken three times a day, while others can be taken once a day or even once a week.

So, and probably also costs to some extent. If, particularly if patients have co-payments or need to pay the, the drug out of pocket, costs probably play a role with respect to adherence and persistence, and since many of these antidepressants are now available as generic medications as well, there really is a broad spectrum of medications that, that can be chosen.

MS. HORN: Fabulous. I think we have just one more question, unless there are more. Again, please use your Q&A function.

“As you mentioned nausea [and vomiting were the most common] reasons for discontinuing medication. Was qualitative research conducted into whether or not a patient would consider trying a different medication?”

DR. GARTLEHNER: Uh huh. Yeah. Nausea and vomiting was the most common reason for patients to discontinue treatment. I'm not aware of any qualitative research that assessed whether patients are willing to switch to another medication, but I think this is, the way to, to go forward. If patients do not tolerate an antidepressant, clinicians need to offer them another choice, and there are antidepressants that have a lower risk for vomiting and nausea than others.

MS. HORN: Great. Well, I think that concludes our Web conference. Thank you so much, Dr. Gartlehner, and Bruce. And particularly you all out there. Thank you so much for your time and we really hope you found this session helpful.

Once, again, we would appreciate your feedback on this session, so please share your thoughts on the short evaluation form that will appear on your screen once the meeting has ended. Again, it should only take a couple minutes. Thank you all, and enjoy the rest of your day.

DR. GARTLEHNER: Thank you.

MR. SEEMAN: Thanks, Christen.