It is humbling that more than 50 years after the introduction of antipsychotics we still struggle with clinical questions basic to their use. Two articles in the present issue, coming from somewhat different perspectives, address one such question, that of antipsychotic dosing.

The article by Wang et al. (1) addresses maintenance treatment in schizophrenia (N=404) by examining different dosing strategies following stabilization of an acute episode with 4—8 mg of risperidone daily. Of note, this study was carried out in China, raising the issue of ethnocultural factors in antipsychotic dosing, a topic addressed in the second article to be discussed (2). After a stabilization period of 4—8 weeks, patients were followed up while taking a dose 1) reduced by 50% at 4 weeks, 2) reduced by 50% at 26 weeks, or 3) unchanged from the previous dose. The times to relapse were 571, 615, and 683 days in the three groups, respectively, while the risks of relapse were 30.5%, 19.5%, and 9.4%, respectively. The study is not without its limitations, including an open-label design, lack of clear criteria regarding the acute episode as well as previous treatment, and rationale for the 50% dose reduction. Notwithstanding the limitations, the findings argue for caution in reducing an antipsychotic dose below that used to control acute symptoms, at least during the first year of follow-up.

The second article, by Gardner et al. (2), also addresses antipsychotic dosing, in this case providing international consensus-based guidelines that speak to dose equivalents and clinical variables that might influence dosing (e.g., age, gender, diagnosis). Their findings underscore the lack of agreement, even among experts, reflected in a 76% consensus for clinical equivalency and 67% for all dosing recommendations. These figures highlight the limitations of such an approach, dependent as it is on individual experience and biases versus an established body of evidence; in point of fact though, empiric data are lacking for a number of the variables examined (e.g., hepatic or renal dysfunction, ischemic heart disease). To their credit, the authors include a number of older, as well as newer, antipsychotics, using 20 mg/day of olanzapine as the reference; they also correctly point out that there is no scientific basis for objective data. For example, imaging studies establishing dose equivalents based on dopamine D2 occupancy, the sine qua non of antipsychotic activity, may represent the "gold standard" (3). However, such an approach faces limitations when one is trying to establish equipotent doses for newer agents, such as clozapine, quetiapine, and aripiprazole, each a departure from the traditional high-affinity D2-blocking compound.

What can we, as clinicians, take away from these lines of investigation? Wang et al. (1) conclude that maintenance antipsychotic treatment, at the same dose as that used acutely, is "more effective in relapse prevention than the reduced dose strategies." This conclusion seems at odds with the longstanding axiom that antipsychotic doses be reduced following acute treatment (4, 5), but it must be placed in context. Notably, the average dose of risperidone used for acute treatment in this study was 4.3 mg/day; thus, the "maintenance" dose in the two groups where it was decreased approximated 2 mg daily. While this dose may be suitable for certain individuals or for certain populations (e.g., individuals in the earlier stages of the illness), it may well be that it is simply too low for a more chronically ill population over the longer term.

It is possible that the recommendation that the antipsychotic dose be lowered is not so much wrong as it is outdated. From the standpoint of antipsychotic use, the "more is better" approach characterized the late 1970s and 1980s (e.g., rapid neuroleptization), giving rise to high-dose strategies (6). However, by the 1990s evidence was accumulating to indicate a lack of clinical benefits with high-dose therapy (7), and this was reflected in subsequent dosing recommendations (8). Further, the field witnessed an increased use of concomitant benzodiazepine therapy for its anxiolytic and sedating properties, diminishing the need to achieve this through increased antipsychotic doses (9). The fact that the average acute dose in the Wang et al. (1) study was approximately 4 mg/day supports this interpretation; indeed, in another study examining risperidone in maintenance treatment the mean modal dose was 4.9 mg daily (10). The shift toward lower antipsychotic dosing does call into question the axiom that doses should be automatically reduced during maintenance treatment, but equally important, it raises the question as to whether antipsychotic doses in acute and maintenance treatment are really that different, especially if we focus on psychosis per se.

For clinicians dealing with schizophrenia, switching antipsychotics is a routine part of daily practice. Comprehensive consensus guidelines, such as those provided by Gardner et al. (2), represent an important clinical tool while reaffirming that much of this work is without strong empiric evidence. By the same token, our decision to modify dosing on the basis of selected clinical variables may be driven as much by intuition as by well-controlled studies. This is not necessarily bad, but it is not science either—one could argue intuitively that increasing antipsychotic doses in the face of partial clinical response is a sound strategy, but data to support this approach beyond a relatively narrow window have not been forthcoming (7, 8).

Much of our attention focuses on the goal of new and better antipsychotics. Given the limitations of available agents, in terms of both efficacy and side effects, we would be remiss in not pursuing this goal. At the same time though, we must acknowledge that there remain numerous unanswered questions regarding antipsychotic therapy and that many of the principles guiding our daily practice are not necessarily based on evidence. Both of the aforementioned articles are valuable for the data they provide, but they also prove useful in reminding us that we still have a great deal to learn about antipsychotic therapy itself. If our goal in care is truly one of improved treatment more immediately, as well as over the longer term, we should ensure a strategy that balances the development of more effective drugs with research directed at optimally using antipsychotics in general. There remains much to learn in each case.