Things were bound to get ugly. The total market for antipsychotics is $10 billion, there are five very similar atypicals jockeying for an extra nibble of that huge pie, and each atypical is backed by an aggressive pharmaceutical firm. You do the math.

Each of the 5 post-Clozaril antipsychotics-Risperdal, Zyprexa, Seroquel, Geodon, and Abilify-has its own quirks, but they are all effective for psychosis, as well as for bipolar disorder. So what is there left to debate? Plenty.

This white collar battleground is found largely in the pages of our psychiatric journals, where several "Big" studies have been published recently. The psychiatric community is now abuzz with some basic questions. Are atypicals really better than conventionals? And aside from obvious side effect distinctions, are some atypicals actually more effective than others?

Please don't call them "Atypicals;" They're now "SGA's.”  

We'll review each of these questions in turn. Before we do, you probably have already gathered that the powers-that-be have declared that "atypicals" shall hereafter be known as "second generation antipsychotics (SGAs)" and conventionals shall be renamed "first generation antipsychotics (FGAs)." After having tried to write articles about antipsychotics using these acronyms and becoming increasingly confused ourselves, we have decided to remain editorial Luddites in the service of clarity--at least for one more issue.

Are atypicals really better than conventionals? The answer depends on whether you are talking about "positive symptoms" (hallucinations and delusions) or "negative symptoms" (flat affect, apathy, asociality, poverty of speech). Atypicals and conventionals are equivalent in treating positive symptoms, but some studies have shown that patients tend to "perk up" more when using the newer agents.

The most recent study to weigh in on this issue was conducted by Veterans Administration researchers with no pharmaceutical funding (Rosenheck, JAMA 2003;290:2693). These researchers noted that although several comparative studies have endorsed Zyprexa over Haldol, these results were influenced by a peculiarity of their study design. Namely, patients on Haldol were not treated with antiparkinsonian agents prophylactically, but only after they started showing EPS symptoms. This, in turn, led to a much higher drop-out rate early in the trials for Haldol. Using the standard LOCF research methodology, the last observation on the rating instrument was carried forward to the final analysis, and all those early Haldol dropouts (due to akathisia or rigidity) unfairly skewed the Haldol outcome averages downward.

In contrast, in this new study, patients were given prophylactic Cogentin, 1-4 mg QD. As a result of this common-sense treatment maneuver, there were no significant differences in dropout rates between the Zyprexa and Haldol groups. And, after 12 months of double-blind treatment, there were no differences in symptom scores between the two groups, either positive or negative.

What about side effects? No big surprises here: Zyprexa caused more weight gain after 12 months (25% vs. 8%), while Haldol, even with Cogentin on board, caused more akathisia (28% vs. 15%). One benefit of Zyprexa was a modest but statistically significant improvement in memory and motor functioning as compared to Haldol.

The bottom line appears to be that atypicals and conventionals are likely pretty equivalent in efficacy when prophylactic Cogentin is used. The apparent benefit of atypicals for negative symptoms is probably an artifact of the higher incidence of EPS with older agents.

Nonetheless, cost aside, no one would willingly choose to be on a medication that requires an anticholinergic to minimize side effects, and which still causes a significantly higher risk of TD. Our verdict: atypicals are still the treatment of choice for schizophrenia, but only if you can afford them.

Are all atypicals of equivalent efficacy? The best way to tell is from head-to-head studies, but unfortunately almost all existing head-to-heads have been sponsored by the manufacturers. Predictably, each sponsored study has managed to finagle the research design and statistics such that the sponsored drug comes out looking a little bit better than the comparator. But if you average studies from different sponsors (and we are talking mainly about Zyprexa vs. Risperdal), the differences average out and disappear.

More confusing are the spate of meta-analyses that have been published over the past few years comparing antipsychotics. The methodology of these pesky documents has become more and more sophisticated (ie, more confusing) with time. The latest entry in the "who's the best at statistics" contest was published in the Archives of General Psychiatry (2003;60:553). The authors (without industry funding) identified 124 welldesigned randomized controlled trials, published or unpublished, that had any bearing on the issue of comparative antipsychotic efficacy.

Their conclusion? That among the commonly used atypicals, only Zyprexa and Risperdal have been shown to be significantly more effective than Haldol. The three newer meds-Seroquel, Geodon, and Abilify-do not have evidence of superior efficacy.

You've probably seen these findings splashed here or there. The usual spin is that "Not all atypicals are created equal." But there are two problems with this study. First, the authors used different study inclusion rules for different medications. For Zyprexa and Risperdal, only studies using doses greater than 11 mg per day and 4 mg per day, respectively, were included. But for the other atypicals, any doses were allowed, including doses of Seroquel as low as 150 mg per day.

The second problem is that the Zyprexa/Risperdal studies were conducted in the old days, when patients entering studies were more treatment-naive and therefore more treatment-responsive. The newer atypical studies were all done quite recently, when it became harder to recruit patients into studies, mainly because they were already being happily treated with Risperdal or Zyprexa in the community. Recruited patients are becoming increasingly treatment-resistant, making it harder to show a robust difference between the newer atypicals and Haldol.

Meanwhile, the few studies that have directly compared the “newer” with the “older” atypicals have shown them to be equivalent in efficacy, casting even more doubt on this Archives Study.

TCR VERDICT: Choosing between Atypicals? It’s the Side Effects, Stupid!