Subscribe to our mailing list

* indicates required

Saturday, January 31, 2015

Can Coffee Prevent Suicide?

Very few drugs can be shown, convincingly, to be associated with a definite reduction in suicide rate. People have tried to show this for antidepressants, but it isn't an easy connection to demonstrate. To my knowledge, only two drugs seem to show a consistent connection with reduced suicide rate: lithium (widely prescribed for bipolar disorder) and caffeine.

 Lithium's tendency to prevent suicide in high-risk groups (like bipolar patients) is fairly well known, and the literature on this is extensive. I'll save that discussion for another time (or you can refer to the Mania chapter in the book I'm writing on mental illness).

The results for caffeine are correlational; they do not prove cause and effect. Still, the evidence is intriguing. In 2011, Michel Lucas et al. published a study in JAMA Internal Medicine that looked at 50,739 U.S. women (mean age: 63 years) who were free of depressive symptoms at baseline (in 1996), then prospectively followed up through June 1, 2006. During 10 years of follow-up (1996-2006), 2,607 incident cases of depression were identified. The researchers went back over the data for coffee consumption. They found that depression risk decreases with increasing caffeinated coffee consumption. What's more, they found a clear dose-response relationship. The relative risk was 0.85 (95% confidence interval, 0.75-0.95) for those consuming 2 to 3 cups per day and 0.80 (0.64-0.99; P for trend <.001) for those consuming 4 cups per day or more. In plain English, that means a risk reduction of 15% for drinkers of 2 to 3 cups of coffee per day and 20% risk reduction for those drinking more than that.

Lucas followed up two years later with an expanded study that looked specifically at suicide risk among coffee drinkers. In this study, Lucas et al. analyzed data for 43,599 men and 164,825 women, covering a period of 20 years for men (14 to 16 years for women), and found a strong dose-dependent inverse association between coffee consumption and suicide, with the risk of suicide 45% reduced for drinkers of 2 to 3 cups per day and 53% reduced for drinkers of more than 4 cups of coffee a day. The researchers corrected for a variety of potential confounding factors, including smoking, antidepressant use, alcohol consumption, age, gender, and marital status. (Read the study to get the full discussion.) Neither tea nor decaffeinated coffee conferred any risk reduction.

The Lucas results confirm two earlier studies. A 1993 study showed lower suicide risk among coffee drinkers in a longitudinal investigation of over 120,000 individuals who were followed for an average of 8 years (Klatsky et al. 1993). Suicide risk decreased monotonically with increasing coffee consumption, and was 80% lower in drinkers of six cups of coffee per day as compared to non-coffee-drinkers.

Likewise, a 1996 study involving 86,626 female nurses found suicide risk was 72% lower among women who drank four cups of caffeinated coffee per day as compared to non-drinkers during the first 10 years of follow-up (Kawachi et al. 1996).

Before you get too excited, bear in mind that a research team in Finland found, in 2000, that suicide risk is 58% greater in people who drink eight cups or more of coffee per day. One can hypothesize that agitation, anxiety, and impulsivity (all of which are likelier if you ingest too much caffeine) may be important effects here. But there are other possible explanations, including the likelihood that people who feel the need to self-medicate with 8+ cups of coffee a day are probably at high risk for suicide for other reasons. Certainly, the evidence is abundant that substance abuse, depression, and suicide are closely linked. Perhaps in a future edition of the DSM we'll see a Coffee Use Disorder diagnosis for people who drink 8+ cups a day.

The studies aren't definitive. We still can't say for sure that coffee-drinking prevents suicide. But it's starting to look as though it's at least possible that the cheapest, most effective, most widely available suicide-prevention drug in the world might well be none other than caffeine.

Buy my books (or click to find out more, at least):

Have you added your name to our mailing list? Thank you. Be well.

Tuesday, January 27, 2015

Do Psychiatric Drugs Take Weeks to Work?

If you've been told by your doctor that antidepressants and antipsychotics take weeks to work, you've been told a myth, and your health care provider clearly either hasn't looked at the actual data for these drugs, or doesn't believe the data.

Alex J. Mitchell, The British Journal of Psychiatry (2006) 188: 105-106, mentions three studies in which effects of antidepressants were measured on short time scales. These were all fairly sizable studies, with 369, 429, or 1,277 patients. In one of the studies, researchers "found that regardless of which antidepressant the patient was taking, there was a measurable early effect on day 1." (Emphasis added.) "By day 3, 20% of patients had shown some improvement, and by day 7 50% had improved. Furthermore, 90% of those who showed any response during the first 3 weeks went on to become full responders. Drug–placebo differences (where apparent) could be detected as early as day 5." In another study, "Patients were requested to complete a self-report mood rating every third day. All patients showed a decrease in depression (and anxiety) within the first 3 days, but with little further improvement in non-responders from days 4 to 6."

The following graph, from Kornstein et al. (2009), shows response data, pooled from four studies, for Lexapro (escitalopram; an SSRI) versus a variety of SNRIs. Patient outcomes were measured using the Montgomery–Åsberg Depression Rating Scale (MADRS). Note the asterisk under the first red square, indicating statistical significance, which occurs at Week 1.

These sorts of graphs are not hard to find and your clinician should be familiar with them. Yet, most psychiatrists and nurse practitioners continue to propagate the myth (and it is just that: a myth) that antidepressants (when they work, which isn't always the case, for all patients) take weeks to show any effect. Certainly, it can take many days for some of these drugs to reach steady-state concentrations in tissues, but the behavioral and mood effects, when they happen, are measurable in as little as one to three days (see Mitchell, referenced above).

The following graph shows that Abilify works for agitation symptoms in as little as 30 minutes (when administered intramuscularly).

Intramuscular Abilify (aripiprazole) causes a reduction in agitation in as little as 30 minutes, although statistical significance is not achieved until 60 minutes (45 miutes for Haldol). From Andrezina et al. (2006), Psychopharmacology 188:3, 281-292.
This is obviously a different situation than depression, but the point is, behavioral effects don't take weeks to occur with these drugs. (Note: Abilify is approved for irritability in autism, treatment of psychosis in schizophrenia, and treatment of depression in an adjunctive capacity.) Abilify, when given intramuscularly in a hospital setting, begins to have measurable effects on behavior in minutes, even though the drug has an incredibly long half-life (50+ hours, or up to 90+ hours for certain metabolites). Abilify is not unique in this respect. Graphs similar to the one above can be found for other antipsychotics here.

So the next time anybody tells you these drugs take weeks to work, understand that you're dealing with someone who either hasn't read or doesn't believe the literature. The literature says effects happen almost as soon as you start to look for them.

I've written a 115,000-word book on mental illness (which is part memoir, part science); it's in the formatting stage now. Please add your name to the mailing list to be notified when it's ready. Thanks so much.

Sunday, January 25, 2015

Do Antidepressants Increase Suicide Risk?

The subject of whether antidepressants increase or decrease suicide risk is a complicated one, and the literature on this subject remains conflicted. A study published in 2004, covering the period 1985 to 1999, found that U.S. suicides decreased 13.5% at a time when antidepressant use rose four-fold. The authors concluded, rather prematurely: "The decline in the national suicide rate (1985-1999) appears to be associated with greater use of non-tricyclic antidepressants." Those results have to be considered invalid, however, because from 1999 to today, suicides (and suicide rates) in the U.S. have increased every year, despite increasing use of antidepressants. The current rate (of 12.6 suicides per 100,000 Americans per year) is higher than before SSRIs were introduced. Are we to conlcude, now, that antidepressants are causing suicide rates to go up? That, too, would be premature.

We know suicide rates increase with age, and the U.S. population is growing older. Therefore, the ever-increasing suicide rate from 1999 to today reflects expected demographic trends.

In "Antidepressants and the risk of suicidal behaviors," JAMA 2004 Jul 21;292(3):338-43, Boston University School of Medicine researchers H. Jick H, J.A. Kaye, and S.S. Jick looked at suicide and suicidal behaviors in a population of 159,810 first-time users of four antidepressant drugs. They found a 4-fold increase in risk of suicidal behavior in antidepressant users in the first 9 days after starting a prescription. That's not the shocker. The shocker is that when they looked at risk of completed suicide in people who had started antidepressant drugs within 9 days, versus people who had been on drugs 90 days or more, there was a 38-fold increased risk for the former versus the latter. (The drugs in question were Elavil, Prozac, Paxil, and the tricyclic dothiepin, which has been sold in various countries as Prothiaden, Dothep, Thaden, or Dopress; it is not sold in the United States)

It's worth noting anecdotally that Robin Williams had started Seroquel just eight days before his death by suicide. He was also taking Remeron, an older serotonergic antidepressant. You can choose to look at his death as a fluke, if you want, or something that would have happened anyway, with or without drugs. We may never know for sure what really happened.

Why the increased risk in the first nine days of treatment? Various theories about this "rollback phenomenon" were discussed in a previous post. Accepted medical wisdom (going back at least 50 years) says that when severely depressed patients are "activated" by treatment, they go from avolitional to volitional and act out preexisting desires to inflict self-harm. This ancient theory needs to be reconsidered, however, as it was formulated back in the days when "clinical depression" was indeed severe and involved psychomotor retardation. Today's user of Paxil, Prozac, etc. is not that kind of patient. The overwhelming majority of "depressive" patients today suffer a less severe depression, often diagnosed by a family doctor rather than a psychiatrist.

"Activation" no doubt accounts for some (small) percentage of self-harm incidents. The other theories that need to be considered involve drug-induced worsening of depression, and/or severe akathisia reactions (which occur roughly 1% to 5% of the time), possibly also coupled with insomnia (which is known to bring a suicide risk in and of itself).

We can argue about the reasons, but the fact is, many people find that with antidepressants, they get worse before they get better. The drugs are no panacea. They work for about a third of users (This was the principle finding of the $35 million, 6-year STAR*D study, and it's also born out by clinical efficacy trials, especially those done in modern times.) In most antidepressant trials, a third of patients drop out of the trial rather than continue on the drug. This too is consistent with the fact that some people find the drugs immediately disagreeable.

If you're suffering, you should consider trying antidepressants, but only do so under close medical supevision; and be aware that many people get worse before they get better.

Buy my books (or click to find out more, at least):

Have you added your name to our mailing list? Thank you. Be well.

Saturday, January 24, 2015

The Curse of Akathisia

A small but important percentage of people who take antidepressants, antisychotics, and certain other meds (like prochlorperazine, used for vertigo, nausea, and migraine) tend to develop a very serious reaction called akathisia, which, in severe cases, is responsible for suicide, homicide, and self-harm.

I urge you to educate yourself on this bizarre condition, because odds are, your clinician is not well trained in recognizing it, and most patients are unaware of its implications.

Older psychiatrists who grew up in the era of "mental institutions" and neuroleptics are familiar with akathisia (properly so-called), but modern scientific literature reflects an incomplete, fuzzy understanding of it, and we see the result in courtrooms across the UK and US. We're caught off guard when a patient with a bad reaction to a drug experiences an abrupt change of personality and kills himself (or his family) as a result of akathisia. Somehow, this happens over and over again, and each case is seen as a fluke, something that's outside the realm of known medicine, when in fact there's a long precedent in history, medicine, and law for this syndrome.

We need to stop thinking of "the guy who went berserk after taking Prozac" (or Zoloft, or Seroquel, or what have you) as mysterious flukes, mere "anecdotal" incidents that defy rational explanation. We need to look at something like Robin Williams' sudden decision to take his life, eight days after starting Seroquel, and ask: Is there precedent for this? Does it conform to anything we know about these drugs? Is it a known type of adverse outcome?

I guarantee that after you, yourself, have either experienced akathisia, or personally witnessed someone having this reaction, you will know that it is a very real thing. It may be rare, but it exists. My wife experienced it and I almost called 9-1-1 several times, until we figured out which medication was causing it (Latuda) and discontinued the med. Within 24 hours, her symptoms went away. I'm capable of having (indeed, eager to have) my mind changed on a lot of things, but no one will ever convince me that akathisia is not real.

Ladislav Haskovec, who first described akathisia in 1901, considered akathisia to be a kind of hysterical reaction, but R. Bing (in a 1939 textbook on nervous conditions) characterized akathisia as a type of "psychosis" involving "morbid fear of sitting down." Hodge (1959) said that akathisia "may appear like an anxiety state . . . in which real anxiety can be neither recognized nor verbalized." Raskin (1972) recognized that patients often are unable to distinguish bctwecn anxiety and restlessness, and warned that "indications of anxiety-like symptoms" such as "uneasiness," hyperactivity, pacing, "vague complaints about medication," and insomnia may actually be reflective of akathisia.

Van Putten, in his classic paper on akathisia, noted that in patients with florid psychosis, the patient may be unable to articulate dysphoric symptoms properly ("making the diagnonsis especially difficult") and can confuse the symptoms of akathisia with those of the original condition for which the patient is being treated. (How, then, can akathisia be reliably diagnosed? For Van Putten, the test was simple: If the patient's restlessness went away with an injection of 5 mg of the anti-Parkinson's drug biperiden, it was akathisia.) Van Putten's patients tell the story better. They experienced fright, terror, anxiety, and rage, and sometimes beat their head against the wall. "I'm frantic. I just can't get my emotions under control. All of a sudden I feel terrified and want to run." "I feel hostile and I hate (with intense affect) everybody." "My nerves are just jumping." "I just feel on edge; I feel nasty; I feel like jumping out of my skin; if this feeling continues, I would rather be dead."

Kalinowsky (Am J Psychiatry. 1958 Oct;115(4):294-8) observed that akathisia can be "more difficult to endure than any of the symptoms for which [the patient] was originally treated." He cautioned that it could be mistaken for "agitated depression."

Fouks ("Le Syndrome d'impatience," 1968) found that akathisia is associated with severe anxiety, peculiar body sensations, and bizarre mentation.

Akathisia reactions are why Thorazine so quickly fell out of use when additional neuroleptics were developed. Subsequent neuroleptics were "better tolerated" largely because they were less apt to induce akathisia.

The antihypertensive reserpine began to replace Thorazine in 1954. But it was associated with an increased rate of suicide—in the hypertensive patients for whom it was prescribed, rather than in the psychiatric patients to whom it was given in higher doses (Healy and Savage, 1998). Its use was associated with "depression," which was induced in at least 10% of patients who took it. The claims for reserpine-induced depression came primarily from physicians rather than from psychiatrists, however, and it is possible non-psychiatrically-trained physicians were seeing akathisia rather than depression per se. Healy and Savage note:
But another state could appear within hours or days of treatment commencing. This was characterised as follows: "increased tenseness, restlessness, insomnia and a feeling of being very uncomfortable" (Achor et al, 1955), "the first few doses frequently made them anxious and apprehensive . . . they reported increased feelings of strangeness, verbalized by statements such as 'I don't feel like myself' . . . or 'I'm afraid of some of the unusual impulses that I have"' (Faucett et al, 1957). Sarwer-Foner & Ogle (1955) describe the case of a subject who on the first day of treatment reacted with marked anxiety and weeping and on the second day "felt so terrible with such marked panic at night that the medication was cancelled."
Such reactions were interpreted by some as evidence for the then-current theory that patients with essential hypertension actually had a suppressed rage. At least one researcher (Ayd, 1958) considered the syndrome "pseudo-depression" and commented on "motor restlessness which made their muscles taut, compelled them to pace the floor and did not permit them to sit without moving their legs."

M. Katherine Shear and colleagues (1983) reported on suicides that had accompanied use of the antipsychotic fluphenazine. Two years later, Schulte (1985) reported on suicide and homicide associated with akathisia. Other reports followed.

With the introduction of SSRIs (selective serotonin reuptake inhibitors; e.g., Prozac, Zoloft), the spectre of akathisia revived anew. Welsh psychiatrist David Healy tells the story of how, in early 1983, almost a decade before it launched in the US, a study of Zoloft (sertraline) was undertaken by Dr. Ian Hindmarch in Leeds, UK, using healthy volunteers. Healthy-volunteer studies are typically part of Phase I trials, to show basic safety. In such studies, non-symptomatic, illness-free volunteers simply take a new drug for up to a week just to show that there are no ill effects. In the Hindmarch protocol, there were 12 female volunteers aged between 34 and 40; the study was supposed to randomize half its subjects to sertraline and half to placebo, for a week, followed by a crossover between drugs. The study was abandoned before the first week was out (and never published). According to Healy:
The medical report to Pfizer noted that the side effects reported in the study were all elicited independently, without communication between participants, that there was a clearcut difference in side effect reporting between placebo and sertraline, and that the volunteers on sertraline were experiencing marked discomfort. The study was accordingly terminated.

All of the sertraline subjects had problems, as had one of the placebo subjects. The placebo subject having problems, however, had sertraline levels in her blood, making the finding even more convincing. The side effects that seemed most clearly linked to sertraline were apprehension, insomnia, movement disorders, and tremors.
Apparently, one of the placebo subjects had just begun the crossover part of the experiment, going from placebo to sertraline; that's when her problems began.

Rothschild and Locke (1991) described three patients who became suicidal on fluoxetine (Prozac), then discontinued the drug, then were reintroduced to fluoxetine—and immediately became suicidal again. Said the researchers: "All three patients developed severe akathisia during retreatment with fluoxetine and stated that the development of the akathisia made them feel suicidal and that it had precipitated their prior suicide attempts. The akathisia and suicidal thinking abated upon the discontinuation of the fluoxetine or the addition of propranolol."

Hamilton and Opler (1992) described a link between SSRIs (specifically, fluoextine: Prozac), akathisia, and suicidality. They noted:
Several reports already exist in the literature documenting the development of EPS [extrapyramidal symptoms] in association with fluoxetine, but without necessarily linking this to an increased incidence in suicidal ideation. Specifically, Lipinski et al. first reported the occurrence of akathisia in five patients treated with fluoxetine. Bouchard et al. reported that EPS developed in several of their patients while they were being treated with fluoxetine and in other patients the baseline levels of EPS worsened during fluoxetine treatment. Symptoms noted included bradykinesia, cogwheel rigidity, and akathisia. Tate reported that a patient who had previously tolerated haloperidol alone had an increase of EPS (including parkinsonism and akathisia) when fluoxetine was added. Stein reported a case of tardive dyskinesia that developed when a low dose of haloperidol was added to fluoxetine. In the case reported by Teicher et al., four of the six patients described complained of an inner restlessness which Opler has previously argued could reflect that they were experiencing akathisia. Wirshing et al. recently reported that five patients treated with fluoxetine experienced ‘agitation, restless motor movement, dysphoria, pacing, an internal sense of desperation, and suicidal ideation,’ and they too suggest ‘that fluoxetine-induced akathisia can lead to suicidal ruminations.’
The literature on akathisia and suicide (and the connection to SSRIs) is fairly extensive and has been summarized by Healy here. In addition, David Healy and Chris Whitaker, in 2003, published an extensive review of suicidality vis-a-vis SSRIs in the Journal of Psychiatry and Neuroscience, in which (quite aside from demonstrating some alarming risk ratios) they tease apart some very important and often-missed issues with regard to the drug makers' tallying of adverse events at various phases (including the "washout" phase) of clinical trials. They point out that 5% of study participants typically report "anxiety" or "agitation" (no patient ever reports akathisia, since it's a term known only to specialists). Also, they note that the drug makers have often customarily, and quite deliberately, prescribed benzodiazepines (e.g., Xanax, Valium, Klonopin) to study participants, to minimize the agitation that's known to occur in some people taking SSRIs. (What's that? You didn't know that drug makers confound their own study results by prescribing other psychoactive drugs concurrently, during trials? Welcome to the real world.)

In antidepressant trials, over the years, side effects involving akathisia have routinely been conflated with "anxiety," "agitation," "restlessness," "insomnia," and other easy-to-score symptoms. (Note that many line employees at the clinical research organizations that carry out these studies are ill trained to recognize akathisia, per se, in study participants.) Combine this with the fact that only a fraction of people with side effects ever bother to complain about them (the true ratio is often considered to be one-in-ten), and you have a formula for disaster.

The literature on suicidality and SSRIs is either clearcut (in favor of SSRIs reducing suicidal behavior) or not, depending on what you read. The extraordinarily exhaustive 2005 literature review by Fergusson et al. in BMJ (which everyone should read) found an increased risk of suicidal behavior, but not completed suicides, for SSRI users. One way to understand such results is to consider the possibility that what we're looking at are two population subsets that produce countervailing results. On the one hand, we have a subpopulation of patients for whom SSRIs work; and these people probably commit fewer suicides. On the other hand, we have a subpopulation of patients for whom these drugs produce a spectrum of "agitation" effects, including (in predisposed individuals) full-on akathisia. The latter commit suicide, or attempt to. Combine the subpopulations, and the net result is an increase in suicidal behaviors, but not actual suicides.

But suicide isn't the only problem.

In a 2006 report in PLoS Medicine, we learn of the case of DS, a 60-year-old man with a history of five prior anxiety/depressive episodes, none of which involved suicidality or aggressive behaviour. His prior episodes had resolved within several weeks. In 1990, DS had an episode of depression, which his doctor treated with fluoxetine (Prozac). The man had a clear adverse reaction to fluoxetine involving agitation, restlessness, and possible hallucinations, which worsened over a three-week period despite treatment with trazodone and propranolol (which should have mitigated such reactions). After fluoxetine was discontinued, DS responded rapidly to imipramine.

In 1998, a new family doctor, unaware of his adverse reaction to fluoxetine, prescribed paroxetine (Paxil, an SSRI), 20 mg daily, for DS, for what was diagnosed as an anxiety disorder. Two days later, having had two doses of the medication, DS used a gun to put three bullets each through the heads of his wife, his daughter who was visiting, and his nine-month-old granddaughter, before killing himself.

At the jury trial in Wyoming in June 2001, a jury found that paroxetine “can cause some people to become homicidal and/or suicidal.” SmithKline Beecham was deemed 80% responsible. The documentary evidence at the trial included an unpublished company study of incidents of serious aggression in 80 patients, 25 of which involved homicide.

Many additional "anecdotal reports" exist. We can add actor/comedian Robin Williams to that list; he had just begun taking the antipsychotic Seroquel eight days before he hung himself in 2014.

These are all "just anecdotes" until they actually happen to someone you know.

Don't kid yourself. Akathisia is real. The body count says so.

This post was adapted, in part, from a book I'm writing on mental illness. Please add your name to the mailing list to be notified when the book and free sample chapters become available. Thank you, and please, if you know someone who can benefit from this kind of information, tell him or her about the book and this blog. Thanks so much.

Friday, January 23, 2015

The Rollback Phenomenon

The debate around suicidality in patients taking antidepressants (which reached a peak a few years ago, when FDA came out with its famous black box warnings) has been a contentious one, with many papers in the literature "proving," with this or that meta-analytical summary, that antidepressant use doesn't bring demonstrable increased risk of completed suicide.

What many people in this debate seem to be forgetting is that for many years, before SSRIs came on the market, it was accepted, as fact, that depression patients do normally tend to show more evidence of suicidality in the early phases of recovery. This was already well known in 1960. If we look, for example, at Slater and Roth’s Clinical Psychiatry, by Bailliere, Tindall and Cassell (London, 1960), p. 231, we find the statement:
“With beginning convalescence (following initiation of treatment with tricyclic antidepressants), the risk of suicide once more becomes serious as retardation fades.” 
This was standard medical-textbook knowledge, until recently. It used to be called the "roll-back phenomenon."

Various mechanisms have been proposed for the increase in suicidality early in treatment of depression:
  • Activation: This view holds that antidepressants with prominent energizing effects may actually increase suicidal behavior in severely depressed patients whose psychomotor retardation was (in the absence of therapy) keeping them in an avolitional state. Untreated, their own stupor/lethargy inhibited such patients from acting out their suicidal thoughts.
  • Paradoxical worsening of depression: This is the view that, in some patients, depressed moods may actually worsen as a direct result of antidepressant treatment. We know that this does, in fact, happen in some patients.
  • Akathisia: This idea holds that some antidepressants produce a side effect of akathisia, which is known to be associated with suicide risk. (This is a complicated subject, which I will address very soon in another post.) Note that akathisia is actually a complex syndrome with mood, mentation, and motor components. It's far more than just "restless legs" or "restlessness."
  • Anxiety: This is the view that certain antidepressants may induce anxiety and panic attacks, which can lead to suicidal behavior in certain patients. (There is reason to believe, however, that many clinicians have recorded a patient's reports of "anxiety" directly, without considering whether the anxiety was actually reflective of akathisia.)
  • Stage shifts: This is the view that antidepressants may cause a switch from depression into mixed states or a bipolar-like condition. Bipolar illness is known to be associated with very high rates of suicidality. The use of antidepressants does bring a risk of mood switch, even in unipolar patients.
  • Insomnia: This view says that insomnia associated with certain antidepressants may lead to suicidal behavior in some patients.

The last point on this list doesn't make much sense until you realize that suicide rates (when looked at by time of day) reach their peak at 2:00 to 2:59 a.m.

I believe good evidence exists for all of the mechanisms proposed above. The "activation" explanation is quite old and has been relied on by many clinicians to explain rollback. However, it doesn't explain the appearance of agitation and suicidality in so-called "healthy volunteer" trials (in which illness-free volunteers take a drug to prove its safety), which are a routine part of FDA Phase I trials.

In my next post, I'll address one of the above mechanisms in more detail: akathisia (which, it turns out, is much more than mere "restlessness").

I'm working on a 115,000-word book on mental illness, part science, part memoir. To stay up-to-date on the book's progress, keep checking this blog and also (please) add your name to the mailing list. Thanks!

Thursday, January 22, 2015

Antidepressant Dose Doesn't Matter

It's common practice, when a patient complains of not seeing improvement on antidepressants, for the clinician to increase the dose of whatever the patient is taking, on the theory that everyone's metabolism is slightly different, and what may be a therapeutic dose for one person may not be strong enough for someone else.

There is no scientific basis for upping the dose of antidepressants, however, since almost all antidepressants have a flat dose-response curve.

If your clinician is playing around with doses, you're dealing with someone who hasn't read the literature. The literature says dose doesn't matter.

Patricia Berney of the Unité de Psychopharmacologie Clinique, Hôpitaux Universitaires de Genève, Chêne-Bourg, Switzerland, did a literature survey on antidepressants and doses, published in Dialogues in Clinical Neuroscience Sep 2005; 7(3): 249–262. Her conclusion:
The results show that a flat dose-response curve is a class phenomenon for selective serotonin reuptake inhibitors (SSRIs), according to randomized, controlled, fixed-dose clinical trials.
It's worth pulling some quotations directly from her analysis of the literature.

For Celexa, she found: "The short-term studies with citalopram did not show significant differences  [...]  When using change on the HAMD and MADRS total score, citalopram 20 and 40 mg/day were no different from placebo."

For Lexapro, she found: "The only fixed-dose-response study with escitalopram indicates that 10 mg/day was equally as effective as 20 mg/day."

For Prozac: "The studies with fluoxetine did not show significant differences in terms of clinical efficacy across a dose range of 20 to 60 mg/day. [...] Therefore, for the majority of patients, there is no advantage of increasing the dose of fluoxetine above 20 mg/day. It might even be the case that the higher dose of 60 mg/day is less effective in major depressive disorder." Also: "The study by Dornseif et al. was performed more than 15 years ago. It is of great importance because it demonstrated that there is no advantage of tripling the dose of fluoxetine to 60 mg/day in outpatients who fail to initially respond to 20 mg/day for 3 weeks; during the next 5 weeks, patients in both groups responded to the same extent and at the same rate."

For Luvox: "Significant differences were not seen between fluvoxamine 25, 50, or 150 mg/day or placebo."

For Paxil: "In the publication by Dunner and Dunbar (Table I), there is a short description of a study involving 460 patients. The paroxetine 10 mg/day dose was no more effective than placebo, even on the HAMD depressed mood item. The authors reported also on a pooled analysis from a worldwide database, involving 1091 patients who remained on a fixed dose of paroxetine or placebo for at least 4 weeks, which showed no differences in terms of clinical efficacy across a dose range of 20 to 40 mg/day paroxetine. Therefore, for the majority of patients, there is no advantage in increasing the dose of paroxetine above 20 mg/day."

For Zoloft: "The SSRI sertraline did not show significant differences in terms of clinical efficacy across a dose range of 50 to 200 mg/day, according to a major study by Fabre and Putman (Table I). Therefore, for the majority of patients, there is no advantage to increase the dose of sertraline above 50 mg/day."

For the tetracyclic antidepressant maprotiline, otherwise known as Deprilept, Ludiomil, or Psymion, versus Paxil: "The study by Benkert et al. used the same protocol as Dornseif et al. and Schweizer et al. and evaluated two antidepressants, paroxetine and maprotiline. This study could not demonstrate an advantage of doubling the dose of paroxetine to 40 mg/day in patients who had failed to respond initially to 20 mg/day for 3 weeks. In another group of 273 patients (not included in Table IV ), no advantage of increasing the dosage of maprotiline to 150 mg/day in patients who had failed to respond initially to 100 mg/day for 3 weeks could be demonstrated. No significant benefits of dose escalation were found."

For Savella, Dalcipran, Toledomin (the SSRI milnacipran, commonly prescribed for fibromyalgia): "In the study by Guelfi et al., milnacipran was prescribed at doses of 100 and 200 mg/day, with a third group receiving fluoxetine 20 mg/day. At the end of 12 weeks, there were no differences between the three groups on change on the HAMD 17 items and MADRS total scores on ITTLOCF."

Ultimately, Berney said:
Our review of eight clinical trials at fixed doses that have evaluated the dose-response relationship of SSRIs in the treatment of major depressive disorders suggests that the dose-response curve is flat. 
Moreover, three augmentation studies could not demonstrate an advantage of doubling the dose of paroxetine, or tripling the dose of fluoxetine, in patients who had failed to respond initially to 20 mg/day for 3 weeks (Table IV). There was a heterogeneity of the results in that some studies did not show a significant difference between the active substance and placebo, or between the highest dosage and placebo.
If you read Berney's report carefully, she mentions a number of side effects for which dose-response relationships were, indeed, noted. Thus, these are drugs for which the primary clinical effect (relief from depression) shows no dose-response curve, whereas side effects do show a dose-response curve. Lack of a dose-response relationship is, of course, typical for placebos.

Berney's results are consistent with a separate meta-analysis by Baker et al.,"Evidence that the SSRI dose response in treating major depression should be reassessed," Depression and Anxiety, (2003), 17(1):1-9. It's also the conclusion reached in yet another meta-analysis by Hansen et al., Med Decision Making January/February 2009, 29(1):91-103 Said the Hansen group: "Dose was not a statistically significant predictor of categorical HAM-D response. Among comparative trials with nonequivalent doses, trends favored higher dose categories but generally were not statistically significant."

These results are consistent with a hypothesis that antidepressants of the types listed above are, in fact, basically nothing more than fancy placebos. Many antidepressants fail to perform better than placebo in efficacy testing, and (just like a placebo) they have no discernible dose-response curve.

If your clinician has been playing around with doses, you're both wasting your time. A placebo is a placebo is a placebo. Dose doesn't matter.

I'm writing a no-bullshit 117,000-word book on mental illness. Please add your name to the mailing list to be notified when the book is ready, or to learn how to get free sample chapters; and please pass this information on to anyone you know who may benefit from it. Thanks!

Wednesday, January 21, 2015

Antidepressant Efficacy: Hard to Prove

Antidepressants have not fared particularly well in clinical trials to determine efficacy. Placebo effect has gotten stronger over the years, resulting in many expensive failures of late-stage (Phase III) drug trials, with the result that most major drug makers have halted R&D programs aimed at developing new psychiatric drugs. Many of today's most popular drugs for depression barely achieved efficacy 30 years ago; if they had to jump through the same FDA hurdles today, most wouldn't make it.

The following table, from a 2003 Neuropsychopharmacology paper by Khan et al., shows the results of fixed-dose clinical trials (done for FDA approval) for a variety of antidepressants: fluoxetine (Prozac), nefazadone (Serzone), venlafaxine (Effexor), sertraline (Zoloft), bupropion (Wellbutrin), and citalopram (Celexa). Note that Serzone has since been pulled from the market over liver toxicity concerns. Trials in which the drug (at the dose shown) was superior to placebo are shown in bold. Remarkably, only a third of trials showed the drugs to be better than placebo.

Antidepressant: protocol
Duration of trial (weeks)
% symptom reduction
% symptom reduction
Dose (mg/day)
1. Fluoxetine: 62


2. Nefazadone: 03AOA-003

3. Venlafaxine: 600A-203


4. Sertraline: 103


5. Bupropion: 203
6. Bupropion: 205

7. Nefazadone: 030A2–0007

8. Mirtazapine: 003–008


9. Venlafaxine: 600A-313

10. Citalopram: 91206


11. Nefazadone: 03AOA-004A

12. Nefazadone: 03AOA-004B

13. Bupropion: 212

14. Citalopram: 89303
15. Venlafaxine ER: 367



How, then, did FDA decide to approve the drugs? Fixed-dose trials are not the only kind of trials used in drug approval. The drug makers also submit flexible-dose trials. Those are successful around 60% of the time. (See the Khan paper for data.)

Fixed-dose trials are typically done to establish dose-response relationships. Since weaker doses might not be as efficacious, it stands to reason that fixed-dose trials (which can be weighed down by weak-dose results) should fare worse than flexible-dose trials in terms of establishing efficacy. In theory, at least. In reality, that's not what Khan et al. found. When they looked at symptom reduction as a function of dose, they were unable to demonstrate a clear relationship ("paradoxically," in the investigators' words). Dose doesn't seem to matter much.

Looking at overall results, the average improvement in patient HAMD score was 35.8% for those who received placebo and 42.2% for those who got an antidepressant. One interpretation of this is that 85% (35.8 divided by 42.4) of the benefit of the drugs is due to placebo effect. But it's also possible that the only thing we're seeing here, in any of the results, is placebo effect. The greater "response" on drugs may be due to breaking of blind, as drug-takers realize (when side effects start to kick in) that they're not in the placebo group and begin thinking they're taking something powerful (when really it's just a different kind of placebo). This would also explain why dosage doesn't matter.

As you know (if you're a regular reader of this blog), I'm writing a mental illness memoir which is undergoing final proofing and formatting right now. Sample chapters will be available very soon. Add your name to the mailing list to be notified when samples are ready. Thanks!

Tuesday, January 20, 2015

Is CBT Better Than Other Therapies?

Of all the "talk therapy" modalities in use today for treatment of common psychiatric disorders (such as OCD, anxiety disorders generally, and depression), probably none is more agressively promoted than Cognitive Behavioral Therapy. As I point out in my forthcoming book, the history of psychiatry is basically a history of fads, and as some fads (e.g., Freudian analysis) fall by the wayside, others (Gestalt therapy, transactional analysis, acceptance and commitment therapy, behavioral activation, and many others) come to take their place. The fad du jour happens to be CBT.

A fundamental tenet of the cognitive model of psychopathology is that cognitive change is central to treating psychological disorder. Modification of maladaptive cognition is core to CBT. The Beck Institute website puts it this way:
When people are in distress, their perspective is often inaccurate and their thoughts may be unrealistic. Cognitive behavior therapy helps people identify their distressing thoughts and evaluate how realistic the thoughts are. Then they learn to change their distorted thinking. When they think more realistically, they feel better.
The goal of CBT is to help the patient learn to identify cognitive distortions and practice reframing them in a more useful way, thereby reducing the patient's reliance on "bad coping skills" like arbitrary inference, selective abstraction, over-generalization, exaggeration (e.g., "catastrophizing"), and minimization. This is a gross simplification of CBT, but it goes to the essence of what gets done in therapy.

CBT has been criticized (appropriately, I believe) on theoretical grounds for its slavishly rationalist bent and its insistence that mentally ill people can fix their problems with what amounts to exercises in clearer thinking. M.B. Shapiro Award winner Chris Brewin (University College London) questions the idea that challenging a person's thoughts leads to changes in feelings and behaviors. He suggests that human cognition actually draws from multiple memory systems and personal-knowledge stores, not all of which are open to introspection. Teasdale (in this book), along the same lines, contrasts “propositional” meanings (which are semantic and declarative) with “implicational” meanings (holistic meanings reflecting the “felt sense” of experience, thus tied to emotion), and questions whether the latter can be assailed with mere logic. Someone suffering from depression isn't necessarily interested in having his logic proved wrong, or being told that every problem can somehow be reduced to a distortion in perception. One of the signature features of mental illness is the convincing illusion that everything you're thinking is 100% real and accurate, even when you're delusional. Indeed, depression is often (subjectively) a feeling of experiencing reality too accurately.

But even in the case of psychosis, it sometimes happens that the patient understands (in his conscious mind, on a logical level) that the FBI isn't really monitoring his brainwaves with listening devices concealed to look like household appliances, yet the feeling that the FBI is "listening" is too strong to ignore. This is the problem CBT can't address (and is why CBT hasn't fared particularly well in schizophrenia trials). You have to be capable of reasoning effectively in order to make use of CBT. In fact, you have to be a high-functioning (not stuporous), motivated, rationality-driven person, and be able and willing to make lists and do homework exercises, in order to get the most out of CBT. You might not be any of the above, if you're profoundly depressed. You might not be any of that if you're not depressed!

Also, sometimes depression is situational, and barring a change in situation (which may not be possible; e.g. if someone died), there's only so much that can be done. If you're grieving over a health situation (cancer), a death in the family, bankruptcy, prison time, etc., you probably aren't ready to have a CBT Pollyanna show you positive-thinking tricks.

In Mindfulness and Acceptance: Expanding the Cognitive Behavioral Tradition (Guilford, New York, 2004), Stephen C. Hayes argues that three “empirical anomalies” exist in the CBT outcome literature. First, component analyses fail to show that cognitive interventions provide added value to therapy. Secondly, CBT is often associated with rapid, early improvement in symptoms, occurring well before any CBT techniques could have been put in practice. Thirdly, measured changes in cognitive mediators (the thoughts and beliefs underpinning the patient's disorder) do not seem to precede changes in symptoms. Richard J. Longmore and Michael Worrell reviewed the literature on these points and found no empirical support for the second criticism by Hayes, but did find substantial evidence in favor of the first and third criticisms.

Cuijpers et al. in 2010 found substantial reason
to suspect publication bias in 175 talk-therapy
trials, most of which were CBT trials.
Notice the asymmetry in this funnel plot.
I won't belabor the issue (since Longmore and Worrell do such a wonderful job on this particular subject), but it's been shown dozens of times, in component factor studies, that the cognitive ingredients of CBT add no demonstrable value (that is, the specific features of CBT can be omitted without adversely affecting outcomes). In fact, more broadly, as Bracken et al. (2012) point out, "The evidence that non-specific factors, as opposed to specific techniques, account for nearly all the change in therapy is overwhelming." According to Bracken:
What emerges from the evidence is that non-specific factors (client variables, extra-therapeutic events, relationship variables and expectancy and placebo effects) account for about 85% of the variance in therapeutic outcomes across the psychotherapy field. In particular, the relationship between therapeutic alliance and outcome seems remarkably robust across treatment modalities and clinical presentations.
Proponents of CBT often point to the massive number of studies showing it to have strong efficacy, but frankly, the studies are subject to many valid criticisms (involving researcher allegiance, lack of a proper placebo arm, and much else), and there is reason to suspect widespread publication bias in this segment of the literature (as in so many others); see accompanying graphic.

Cognitive Behavioral Therapy is effective for many people. So are other therapies. The issue isn't that CBT has no value, but that it's been oversold, and isn't the right tool for every job. CBT can't be said to be superior to other treatments for depression. It's relatively ineffective for schizophrenia and not well supported for bipolar. Even when it works, the effect size is highly variable. See this meta-analysis by Lynch, Laws, and McKenna (2010) for details.

I'm finishing a 113,000-word mental illness memoir that has a lot to say about psychotherapy, drugs, depression, mania, schiophrenia, schizoaffective disorder, and recent trends in psychiatry. To find out more about the book and how to get free sample chapters, add your name to the mailing list and return to this blog often. Thanks!

Sunday, January 18, 2015

Are Psychiatric Drugs as Effective as Other Drugs?

Ever since Irving Kirsch published his famous meta-analysis of FDA clinical trials on antidepressants, showing that half of trials produced no separation from placebo and that antidepressants are generally of dubious clinical worth except for the most severely depressed patients (a result that was consistent with another famous meta-analysis, published in JAMA), various apologists for the pharma industry have tried to spin the results in a more drug-favorable light. One of the most interesting counter-spin papers is the 2012 report by Stefan Leucht and colleagues in The British Journal of Psychiatry.

The Leucht paper can hardly be considered unbiased, since the lead author received fees for consulting and/or lectures from Bristol-Myers Squibb, Actelion, Sanofi-Aventis, Eli Lilly, Essex Pharma, AstraZeneca, MedAvante, Alkermes, Janssen/Johnson & Johnson, Lundbeck Institute and Pfizer (and grant support from Eli Lilly), while another author received fees for consulting and/or lectures from Janssen-Cilag, Sanofi-Aventis, Johnson & Johnson, Pfizer, Bristol-Myers Squibb, AstraZeneca, Lundbeck, Novartis and Eli Lilly. Still, it's an interesting paper and deserves to be read in detail.

SMD (a proxy for effect size) for studies of non-psychiatric
medications (white circles) versus psychiatric medications
(filled circles, right).
The paper attempts to put criticisms of antidepressant effectiveness in context by showing that when studies (and studies of studies: meta-analyses) of antidepressants are stacked up against studies of "regular" medicines (for somatic conditions), antidepressants don't come out looking so bad. To illustrate their case, the authors prepared the accompanying chart showing effect sizes (actually: standardized mean differences) for various medical studies (white circles) versus studies involving psychiatric drugs (solid circles).

What the authors basically tried to show is that if you transform a big bunch of heterogeneous studies into SMD form (mostly using "responder" data, not remission or cure rates), and don't put confidence intervals on the dots, you find that psychotropic drugs are, on the whole, not inferior to other drugs.

However, if you dig into the data (see Table 2 of the Leucht paper), you find that psychiatric drugs have the strongest data in a comparison of drug versus placebo for relapse rates rather than primary treatment effects. The highest blue dots on the graph are from this kind of data. The fact that people relapse in huge numbers when they withdraw from drugs is thus shown in the graph as a strength for drugs, when in fact it's arguably a weakness. (Relapse rates are known to be higher for drugs, for example, than for psychotherapy.) But even if you don't back out some of the high points, response rates for antidepressants are generally unimpressive, with the ratio of drug response to placebo response being 53.3% to 42.4% in one case, 46.0% to 31.0% in another case, and 37.8 to 24.2% in the third of the three analyses looked at. When you combine this with the fact that 18% to 36% of drug responders eventually relapse, the overall picture can hardly be described as encouraging.

I was also struck by the fact that response rates to antipsychotics, for schizophrenia patients, is quite unexpectedly poor. Here, the only two studies cited (which were by Leucht and colleagues) showed a response rate of 40.6% for second-generation antipsychotics (versus 23.7% for placebo) and 29.3% for Haldol (versus 19.5% for placebo). The low response rates here may reflect Leucht's (admirable) tendency to rely on more recent meta-analyses. Early (and generally smaller) meta-analyses of nearly all the psychiatric drugs showed much larger effect sizes than more recent studies.

Speaking of placebo, the Leucht review offers a striking demonstration of the power of the placebo effect, showing, as it does, that placebo response is significant even for diabetes (21.7% for metformin, the only drug reported), stroke (55.8%), breast cancer (42.4%), and antibiotics (22.2% to 25.7%). No doubt spontaneous remissions account for some of the cures reported for "placebo"; but the same is true in mental illness (which is notoriously episodic, regardless of category).

It would have been quite telling if Leucht had produced a graph (similar to the one above) showing only placebo response rates. The graph would look quite similar to the one they produced (albeit with dots shifted slightly lower). Such a graph would demonstrate that placebos for psychiatric trials are non-inferior to placebos for other drug trials.

Of course, one problem with analyses of the Leucht kind is that they're 100% biased in favor of published results. We aren't seeing results from studies that showed no efficacy, because such studies aren't generally pubished very often. This is an important omission, because we know from the work of Kirsch and others that FDA is sitting on a ton of clinical trials that showed safety, but not efficacy, for a variety of psychiatric drugs; indeed, this was a major finding of the Kirsch meta-analysis (see link at the top), which drew from published and unpublished FDA trials obtained by Freedom of Information Act request. Kirsch found that half of antidepressant trials submitted to FDA failed to establish meaningful separation from placebo (and those trials rarely got published, or if they did, their data was conflated with data from more successful trials). To learn how drug companies hide data and spin "lack of efficacy" into efficacy, read the excellent paper by Spielman and Parry (2009), "From Evidence-Based Medicine to Marketing-Based Medicine: Evidence from Internal Industry Documents."

Leucht et al. claim, with their paper, to have put psychiatric drugs in the proper medical perspective. To some extent, this is true. On the whole, though, it doesn't change the fact that drugs for serious mental illness (schizophrenia, bipolar, depression, etc.) are ineffective (producing no response, or even a negative response) for the majority of people who try them, a result that's clear from Table 2 of the Leucht paper.

I am 99% done writing a mental health book (112,000 words) that's part science, part memoir. It's not an anti-psychiatry screed, although it's (appropriately) critical of many aspects of the profession. The 275 footnotes will let you decide which parts are more (or less) believable than other parts. To learn more about the book, add your name to the mailing list and return here often. Thank you!

Saturday, January 17, 2015

The 31% Solution

In writing a book on mental illness (which is topping out, now, at 113K words), I find that although I haven't exactly written an anti-psychiatry screed, I've come about as close to that designation as you can come without sounding like a full-on psychiatry-bashing lunatic. Where's the dividing line? For one thing, I stop short of saying there's no such thing as mental illness; I believe that the utility of the illness model is (just barely) useful enough not to discard it completely. I also stop short of saying psychiatry has nothing to offer, or that drugs have nothing to offer, whereas many anti-psychiatry zealots are willing to go there. I can't go there unless that's where the evidence points. We know the evidence doesn't 100% point there, so let's not pretend it does.

On the other hand, psychiatry (not only psychopharmacology, but modern psychotherapy) deserves criticism on many fronts, and it's not hard to produce evidence on many of those points, so that's all fair game. My book is full of criticism. (And footnotes. Hundreds of footnotes.)

In medicine, there are many "31% solutions," which is to say therapeutic options that work for about 31% of people who try them. This includes placebos, antidepressants, antipsychotics, herbal cures, cures for alcoholism (the year-over-year retention rate for Alcoholics Anonymous is about 31%), various psychotherapies, low-dose electroconvulsive therapy, sham ECT, trans-cranial magnetic stimulation, and probably diets, exercise, and much else. Frankly, a 31% response rate, whether it's for a cancer treatment or a mental disorder, whether it involves "real medicine" or placebos or sham treatments, is not something you just walk away from. When people are desperate for relief, they'll try something that works 31% of the time. And well they should!

But we should be honest about response rates, remission likelihoods, etc., in treatment of mental illness, as in treatment of cancer or any other malady. And this is where the "mental health industry" (if you can call it that) has done a remarkably poor job. Patients are led to believe antidepressant drugs mostly work, when in fact they mostly do not. They've also been told ECT (electroconvulive therapy) mostly works, when in fact the benefits beyond the one-month point are essentially nil and there's no difference in long-term outcome between sham ECT (placebo) and real ECT, something I go into it the book (citing FDA's own internal literature review).

Also, CBT (cognitive behavioral therapy) has been so grossly oversold as a treatment modality, with claims of its superiority to other modalities, that I felt an obligation to counter some of the misleading claims made for CBT. So in the book, I devote a few pages to that. Again, not because CBT doesn't work for some people (it clearly does) or doesn't have any utility at all, but because it's been preposterously oversold. I'll devote a future blog post to that.

For more information about my upcoming book and how to get free sample chapters, please add your name to the mailing list (no updates have been sent out yet; I'm not interested in spamming anyone). And please check back here often. Thank you!

Friday, January 16, 2015

Can Antidepressants Make You Bipolar?

Bipolar disorder is a spectrum disorder characterized by at least one manic or mixed episode, alternating (typically) with depressive episodes. It conforms to Kraepelin's original concept of manic-depressive illness, which can include psychosis symptoms during mood episodes. In bipolar I, you see swings between manic and depressive episodes; in bipolar II disorder the individual never experiences a full manic episode but instead exhibits less severe hypomania (which has the expansiveness of mania but lacks the delusional grandiosity). There's also a somewhat controversial diagnosis of Cyclothymic Disorder, which refers to a serious lability of mood that (for whatever reason) fails to meet the diagnostic requirements of bipolar.

On Showtime's Homeland, Claire Danes (left) played a bipolar CIA agent.
(Mandy Pitinkin, right, played opposite her as a CIA higher-up.)
Despite bipolar's depiction in modern culture as being somewhat common, it's actually quite rare. Bipolar disorder occurs naturally (in the U.S.) at a lifetime rate of 4%. That figure turns out to be quite a bit higher than the condition's actual worlwide prevalence (rate of actual occurrence at any given moment). The World Mental Health Survey initiative found a prevalence of bipolar II in the population of eleven countries to be 0.4%, making it less common than schizophrenia. The prevalence rates of bipolar I and II are thought to be comparable. In the U.S, each occurs with a prevalence of about 1%.

The Diagnostic and Statistical Manual of Mental Disorders (5th Ed.), in its formal criteria for bipolar, discounts mania as mania (or as hypomania) if it's induced by drugs, which is extraordinarily convenient, in that antidepressant drugs can and do cause hypomania in some people. I say "convenient" because the drug companies are keenly interested in dispelling the notion that (unipolar) depressed patients who start taking antidepressants are often, in so doing, converted to bipolar patients, which is exactly what does happen at several times the spontaneous rate.

In a literature review of 51 studies involving nearly 100,000 patients with MDD without a history of mania or hypomania who were treated with an antidepressant, Baldessarini et al. (2013) found bipolar mood switching occurred in 8.2% of patients within an average of 2.4 years of treatment. They also said: "We considered 12 studies of patients initially thought to have unipolar MDD who required a new diagnosis of bipolar disorder, usually with verification by occurrence of spontaneous mania or hypomania. Such diagnostic changes occurred in 3.3% within 5.4 years, or 5.6 times less than the rate of mood switching with antidepressants."

What it means is that if you go to your doctor with major depressive disorder (unipolar depression) and start taking antidepressants, there's an 8.2% chance you'll be bipolar within 2.4 years. Some people react to antidepressants by going manic or hypomanic. It may not happen in the majority of cases, but it is a significant chance, and the longer you're on antidepressants, the greater the chance of a changeover to bipolar.

From personal experience, I can tell you that if your daily routine includes coffee in the morning and alcohol at night, you stand a much better chance of avoiding the bipolar switch if you cut alcohol out of your life completely and keep coffee and other stimulants under control. Coffee and alcohol are cyclothymia triggers (potentially). You want to moderate their effects.

What happens if you do experience mania? Is there a known-good way to treat it? Actually, there is. It's called lithium carbonate, and it's one of the best proven, least expensive anti-mania treatments money can buy. It's also fast-acting and can be taken "PRN" (which means as needed), although it's rarely prescribed that way. The problem with lithium is that today's doctors are using 50-year-old dosing recommendations that often result in a patient taking toxic amounts of the stuff. A typical dosing regimen is 450 mg two to three times a day. That kind of dose will make you toxic in a couple of weeks, but unfortunately, there are few modern studies involving lower-dose lithium. One such study that's worth considering is Alevizos et al. (2012), which looked at the response to adjunctive low-dose lithium in 47 severely depressed patients who failed to respond to venlafaxine (Effexor) alone. Said the researchers: "After 5 weeks of augmentation, 51% of the patients were rated as 'much' or 'very much' improved. Bipolar patients showed a better response than unipolar (64.3% vs 45.5%, p<0.038). Most patients (76%) showed a rapid response (up tp 7 days)."

Two things need to be mentioned with regard to the Alezivos study. First, there was no placebo control arm, hence the results await validation by a properly controlled, properly blinded study. Secondly, "low dose" here meant 300 to 450 mg per patient per day (corresponding to about 5 mg/kg), which is still a high dose. Modern medical science still hasn't wrapped its head around the idea that you can benefit from low doses of lithium, even though we have tantalizing evidence (see, for example, this study; there are others) that micromolar concentrations of lithium in tap water can prevent suicide.

Much of the information in this post was taken from a 110,000-word book I'm writing on mental illness. To follow the book's progress, add your name to the mailing list and return to this blog often. Thanks!

Wednesday, January 14, 2015

Do Mood Stabilizers Work for Depression?

Because modern antidepressants so often fail to give people relief, it's become fashionable in recent years for prescribers to add an adjunctive mood stabilizer (such as Lamictal, Depakote, lithium, valproate, or Tegretol) into a patient's antidepressant mix, yielding a so-called "cocktail" aimed at producing better results. This is a particularly popular approach with bipolar illness (where mood stabilization is critical), although it's also increasingly used in treatment of unipolar depression.

I decided to take a look at the scientific literature to see what evidence there is that such cocktails work. I almost wish I hadn't.

I was stunned to find that the evidence on adjunctive mood stabilizers is both scanty and low-quality, with studies often unblinded as to assessment, open-label (unblinded as to the patient's knowing what he's taking), with small study populations, and producing no clinically significant effect.

One study of lamotrigine (Lamictal, a popular, well tolerated anti-seizure medication with mood-stabilizing side effects) involved just 14 people. Another involved 31 patients. A study with 40 patients found "Adjunctive treatment with lamotrigine did not result in a significant difference in HAM-D total score at the endpoint of the study when compared with paroxetine alone." Another study, with just 20 patients (and no placebo arm), found "there was no significant difference between treatment groups." Another (a lamotrigine monotherapy study) looked at depression-score improvements in 32 epilepsy patients (and 38 controls); this was, at least, a randomized, double-blind, placebo controlled study. And it did find significantly better improvement on Beck Depression Inventory scores, after 12 weeks, for the lamotrigine group (8.9 point drop) versus the placebo group (1.7 point drop), but two things need to be noted about this otherwise-good study. First, the patients (with an average Beck score of 18, at the beginning) were not very depressed. (Moderate depression is 19 to 29 on the Beck scale.) Secondly, when your epileptic seizures go away (which they did tend to, in this study) your mood naturally tends to lighten quite a bit. The investigators claimed to have corrected for this. I don't believe it for a second. You'd have to be crazy to count this study as "proof positive" that lamotrigine works for depression (in non-epileptic patients, particularly).

When I eventually stumbled onto a double-blind, placebo-controlled study involving 350 bipolar patients, published in The New England Journal of Medicine, I thought I'd hit the jackpot. This study, by Sachs et al. (2007), seemed like it would finally shed light on the whole mood-stabilizers-plus-antidepressants issue. But it turns out to be a real mess, one of the shoddiest, muddiest piece-of-crap studies I've ever seen in NEJM, which many consider to be a top-flight journal.

Even if we overlook the conflict of interest issue, namely the fact that most of the paper's 20 authors have received "consulting" and speaking fees from Abbott Labs, AstraZeneca, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, Merck, Pfizer, Sanofi, and Wyeth (indeed, many individual authors received fees from every one of these companies), we're left with a horribly complex study design in which patients received either bupropion (Wellbutrin) or paroxetine (Paxil) as an antidepressant, or a placebo, combined with a mood stabilizer consisting of lithium or valproate or a combination of lithium and valproate or carbamazepine (Tegretol); and then, in 2004, "the protocol was amended to define mood stabilizers operationally as any FDA-approved antimanic agent." When you look to see what the latter means, it means (for 8% of one group and 11% of another group) "atypical antipsychotic drugs" that are never named! Nowhere in the study are we told the identity of the "atypical antipsychotic drugs," no doubt to avoid embarrassment of sponsors.

Even though a third of people left the study (and another 7% had worsening of symptoms; think about that), there is no intent-to-treat analysis, per se. We're merely advised that "Except where noted, analyses are based on the last observation carried forward."

The study ran 26 weeks (which is fairly long) and used a stringent outcome measure: 8 weeks of euthymia (remission, essentially). The measurement tool for this? "A standardized clinical monitoring form adapted from the mood-disorder modules of the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, fourth edition." Except, after you adapt something, in what sense is it now "standardardized"?

Parsing this study only makes me crazy, so let me cut to the chase and mention the principal finding, which was that 23.5% of patients receiving a mood stabilizer plus antidepressant therapy had a durable recovery, whereas  27.3% of those receiving a mood stabilizer plus a matching placebo met remission standards. (Yes, you read that right: the placebo group outperformed the treatment group. And yes, they considered an antidepressant to be the adjunctive drug, not mood stabilizers.) What they really found, of course, is that neither Wellbutrin nor Paxil is efficacious for bipolar illness when measured against lithium, valproate, lithium plus valproate, Tegretol, or unnamed atypical antipsychotics.

Naturally, the authors try to spin the results as positively as they can, but their Conclusion says it all:
The use of adjunctive, standard antidepressant medication, as compared with the use of mood stabilizers, was not associated with increased efficacy or with increased risk of treatment-emergent affective switch. Longer-term outcome studies are needed to fully assess the benefits and risks of antidepressant therapy for bipolar disorder.
Longer-term studies? Longer than 26 weeks? Are you kidding me? Seriously?

Goodwin et al. (2004) did find, in a pooled analysis of two 18-month double-blind studies of lamotrigine monotherapy (versus placebo) that lamotrigine (Lamictal) was superior to placebo in extending the time between interventions for depression, but cites "median survival: placebo, 270 days [95% CI = 138 to not calculable]; lithium, median not calculable; lamotrigine, median not calculable." So I'm not sure what their conclusion is based on, and anyway, who cares about times between intervention? What about obviating the need for an intervention?

I haven't done an exhaustive, systematic review of the literature. However, Ryan J. Van Lieshout and Glenda M. MacQueen (BJPsych, 2010) did do such a review. Because of heterogeneity of study designs, this systematic review deserves to be read in detail. The authors did find that mood stablizers (in the 10 studies that were looked at for clinical response) produced a signifcant effect relative to placebo in monotherapy (no antidpressant used); risk ratio was expressed as 1.3. For adjunctive use-cases (antidepressant plus mood stabilizer), across six studies involving 1,322 participants, combination therapy was not superior to monotherapy (RR= 0.85).

You can find good quality studies that show a beneficial effect of lamotrigine in monotherapy for bipolar depression; but in bipolar, you're looking to stabilize mood (eliminate or reduce mood-switches), and so outcomes are often measured on that basis (rather than on a basis of treating depression, per se). Extrapolating bipolar results to unipolar depression is thus (in my opinion) ill advised. Very few good-quality studies (in fact, none that I could find) specifically address the adjunctive use of mood stabilizers in treatment of unipolar depression. The systematic literature review by Vigo and Baldessarini (2009) on this subject found "suggestive evidence of beneficial effects of carbamazepine, lamotrigine, and valproate that require further study, especially for long-term adjunctive use." I think that probably sums it up. The evidence is suggestive, at best, and the whole matter needs further study.

The science I saw was shoddy, scant, and unconvincing. Much of it should never have been published in the first place.

I'm 100,000 words into a book about mental illness—part science, part memoir—based on my experience with bipolar illness and my wife's experience with schizoaffective disorder. Please add your name to the mailing list to be notified when free sample chapters are ready. And if you know someone who may benefit from this kind of info, please pass the word along. Thanks so much!