The STEP-BD study, an NIMH-funded study of bipolar disorder, has enrolled its last patient, and we have already begun to see the first results filtering through to the psychiatric journals.

STEP-BD stands for “Systematic Treatment Enhancement Program for Bipolar Disorder,” and it is the latest installment in a recent trend in psychiatric research known as “effectiveness studies.” These studies have in common an effort to enroll large numbers of real world patients, using few exclusion criteria, in order to make the results more generalizable to our practices. In prior issues, we have reviewed two other effectiveness trials: the CATIE trial (TCPR Feb 2006) and the STAR-D trial ( TCPR May 2006).

The STEP-BD study enrolled a total of 4,360 patients. In order to be enrolled, patients merely had to meet criteria for bipolar disorder I, II, NOS, or cyclothymia. They could be teetotalers or substance abusers, high-powered executives or chronically homeless. There were essentially no exclusion criteria.

On study entry, patients were assigned a “STEP-BD-certified” psychiatrist. These psychiatrists received 20 hours of training on best practice procedures for how to treat patients with bipolar disorder. These “best practices” were written by a committee of bipolar disorder experts, who identified nine different clinical decision points corresponding to nine common clinical situations in the treatment of bipolar patients. At each decision point, there is a defined Standard Care Pathway with a “menu of reasonable choices.” This is not really a strict treatment algorithm, but rather a series of loose guidelines meant to point treaters in the appropriate directions.

Patients were seen every three months for the first year, then every six months thereafter. They were encouraged to stay in the trial for five years.

The first three major STEP-BD findings

1.Even gold-standard treatment isn’t terribly effective. One of the first papers focused on 1469 patients who had been followed for two years (Am J Psychiatry 2006;163:217-224). After up to two years of best practices treatment, 58.4% of these patients met criteria for full recovery, defined as not meeting more than two DSM-IV criteria for depression or mania for eight consecutive weeks. Note that this is a much more stringent outcome measure than is typically used in controlled trials, in which “response” is measured as a 50% reduction in symptoms at a single time point. The group of 858 patients who recovered by STEP-BD criteria were observed further, and 48.5% of them suffered a recurrence at some point during the study (416 patients), more commonly to a depressive episode (72%) than to a manic, hypomanic, or mixed episode (28%). Not surprisingly, having residual symptoms was associated with a greater chance of recurrence.

What are the practical implications of these data? Patients who present to a bipolar disorder specialist in an acute episode of bipolar disorder (depressed or manic), have slightly better than a 50/50 chance of recovering in two years. Once recovered, there is a 50% chance of relapsing. This may not sound very heartening, but Gary Sachs, STEP-BD’s lead investigator, takes a more optimistic view of the data in this month’s interview (pages 5-6).

2. For treatment-refractory bipolar depression, Lamictal and Inositol beat Risperdal. The STEP-BD study randomly assigned treatment-refractory patients with bipolar depression to one of three adjunctive treatments: lamotrigine (up to 250 mg QD), inositol (up to 25 g QD), or risperidone (up to 6 mg QD) for up to 16 weeks (Am J Psychiatry 2006;163:210-216). All patients were already on a mood stabilizer (lithium, Depakote, or Tegretol) and one or two antidepressants. The recovery rates were: Lamictal, 23.8%, inositol, 17.4%, risperidone, 4.6%. While these numbers may look different, the differences aren’t statistically significantly because of the small number of subjects. Nonetheless, it’s a bit humiliating for Risperdal, which could not outperform inositol, an isomer of glucose that is used by our neurons transmitting information from neurotransmitters.

3. Yes, Depakote really does cause polycystic ovariansyndrome. It’s a notion that has been around for a while: Depakote may increase the risk of developing polycystic ovarian syndrome (PCOS). While a prior study of bipolar women suggested that this happens, the sample was not high enough to yield a statistically significant result. The large sample size of STEP-BD allowed investigators to evaluate 230 women (ages 18-44) for the development of symptoms of PCOS (menstrual irregularities, hirsutism, acne, male-pattern hair loss, elevated testosterone). They found PCOS symptoms in nine of 86 women on valproate (10.5%) as opposed to only two of 144 women on a nonvalproate anticonvulsant or lithium (1.4%). The upshot? Warn woman about these potential side effects before starting them on Depakote (Biol Psychiatry 2006;59:1078-1086).

TCPR Verdict:
STEP-BD -- the findings are rolling out.