Dr. Sachs: I don’t think that by itself is impressive. What is impressive, however, is that the methods used in the study allowed us to be very clear about the starting point of each patient. So one important aspect of the data reported by Perlis et al is that we have a defined starting line. The relapse rates were based on a group of 858 patients who were definitely in remission at the beginning of the observation period. We know this is true, not because we asked them a bunch of questions on one given day, but because we actually observed them and charted their mood events closely from a point when they were acutely ill. Compared to otherstudies,thisprovidesrobustprospectivedataaboutthecourseofillnessafterremission. Weestablishedtheonsetofrecovery by observation, and that has not been done on such a large cohort before.

TCPR: All right, so you’re saying that the quality of the data is unusually good, making the study itself impressive. But doesn’t that make the response rates even more discouraging? I mean, here we have patients being treated by psychiatrists who took a special course in best-practice guidelines, and still only 58% of patients got better after two years of treatment.

Dr. Sachs: It’s important to realize that the primary outcome we used in STEP-BD, achieving a durable remission, required patients to be well for at least eight consecutive weeks. That is a very important goal, and in patients’ eyes is absolutely minimal, but much more stringent than what is typically used in studies. Our durable remission criteria would be the least amount of doing well that anybody might care about, but it is eight times longer than what is used in most studies. Most studies would categorize recovery as only one week of 50% improvement on a rating scale.

TCPR: So STEP used an unusually stringent definition of recovery, making it difficult to compare the results with other studies.

Dr. Sachs: Yes. Similarly, most studies in this field limit enrollment to the kinds of patients that don’t exist. STEP involved real- world conditions. Participants were not symptomatic volunteers who responded to researchers’ recruitment posters and then met narrow inclusion criteria. They were treatment-seeking patients who had comorbid anxiety disorders, substance abuse, and a variety of other problems, making the outcome pretty encouraging.

TCPR: Are there any other findings of note from the Perlis study?

Dr. Sachs: One important clinical pearl from the study is that there is a three-to-one ratio of depression to mania among relapsers. In bipolar disorder, the name of the game is relapse to depression, which helps you understand why there was so much emphasis on bipolar depression in STEP. Even though mania is the cardinal diagnostic feature for bipolar disorder, it is not the biggest problem that a bipolar patient encounters.

TCPR: Another interesting STEP study was recently published, comparing three different treatments for treatment-refracto- ry bipolar depression (Am J Psychiatry 2006;163:210-216).

Dr. Sachs: Yes. It is the only randomized trial that has ever been done with people who had well-defined treatment-refractory bipolar depression. It is important, even though it is small, because it really is unique.

TCPR: But the recovery rates are low: Lamictal (lamotrigine), 23%; inositol 17%; and Risperdal (risperidone) 4.6%. Again, convince us that this is something to get excited about.

Dr. Sachs: These were not just patients with bipolar depression who are already hard to treat, but refractory bipolar depressed patients, who were taking a mood stabilizer plus one or two antidepressants at study entry. And in order to qualify as “recovered,” they had to be completely well for eight consecutive weeks. So this study really is meaningful – these are very refractory patients, and the outcome for lamotrigine is actually a pretty healthy response rate. It is interesting that inositol seems to do reasonably well, but when we started the study, a lot of bets were on risperidone. There was a belief in the late 1990s that the atypical antipsychotics all had antidepressant properties, but, at least in this case, risperidone did not. The bottom line is that failing to respond to a standard antidepressant does not necessarily mean that you can’t recover from bipolar depression. We really have a pretty good response rate, with one in four having a very good outcome with lamotrigine.

TCPR: However, not only was this study open label, but there was no placebo arm. The potential criticism is that if some patients had been on placebo, they might also have achieved a response rate in the 20% range, making all the active treatments look ineffective.

Dr. Sachs: Placebo-controlled studies are certainly ideal, but they are difficult to conduct. Furthermore, the thought was that since nothing had ever been shown to work for these kinds of patients, letting them suffer with a treatment that we had no reason to believe would work (placebo) didn’t make a lot of sense.

TCPR: Why would inositol be almost as effective as Lamictal? Inositol is a sugar.

Dr. Sachs: Inositol is a sugar, but it is involved in the signaling cascade through G proteins, inositol monophosphate, and inositol triphosphate. This is a pathway that lithium affects quite a bit. So there were theoretical reasons to try it, as well as some small studies that found it promising.

TCPR: Thanks for your time, Dr. Sachs. I look forward to more interesting findings from STEP-BD over the coming year.