The schizophrenia Algorithm

In Reply: We thank Drs. Berman and Fleishman for their comments, and note with interest that their letters define the spectrum of questions about clozapine: is it a first-line medication, or the drug of last resort? First, to consider Dr. Berman's question: clozapine has the best outcome data; why not use it first?

In the schizophrenia algorithm of the Texas Medication Algorithm Project (TMAP), all roads lead to clozapine. Our rationale for this course is contained on pages 34-35 of our physicians' manual (1), available on the Internet (2). In brief, improvement in negative symptoms, clozapine's strong point, is a complex event perhaps involving positive symptom relief, the absence of extrapyramidal symptoms, and the absence of depression. This result may be achieved by other medications with side-effect profiles generally thought to be more benign than that of clozapine. In the TMAP algorithm, the response to other medications is evaluated in six to eight weeks, a period far shorter than the three to six months it takes to evaluate response to clozapine. If clozapine does not work, stopping it can be arduous, often producing a worsening of psychosis. Crossover is probably easier with other medications.

The TMAP algorithm contains criteria for improvement that are more stringent than those used in the seminal clozapine study (3) and, to the best of our knowledge, are the most stringent criteria yet developed describing acceptable response (Miller A, Chiles J, Chiles J, et al, unpublished paper, 1999). In the TMAP algorithm, patients who do not improve to the extent that they have only mild symptoms in a few areas will move on to the clozapine stage.

Having said that, we remain most interested in evaluating clozapine as a first-line drug. We are aware that a large study in China is currently evaluating this question (Lieberman J, personal communication, 1999). The TMAP schizophrenia algorithm is a dynamic instrument that has been revised four times in three years. As our knowledge expands, we expect future revisions.

Dr. Fleishman asks if we enter patients into the algorithm who are doing well on their current regimen: we do not. Patients are entered only if their physician decides to change medication, a decision based on either symptoms or side effects. As to why clozapine is listed at stage 5 as an option before a combination of medications is tried, we believe the current literature best supports this monotherapy approach. We are quite aware of the increasing case-report literature on combinations of antipsychotics, and, as noted above, the algorithm can change.

Last, as use of the TMAP algorithm increases, we have found a different issue with clozapine to be perhaps more pertinent: physicians are reluctant to use it, even when their patients have an inadequate response to other treatments. The reasons are not clear, but dread of agranulocytosis seems to be at the forefront. Some percentage of patients with schizophrenia—in our estimation, 10 percent—should be on clozapine. We are aware of a variety of databases that show usage to be nowhere near that level. Physicians themselves need to advocate the use of this drug.