Is the Patient’s Diagnosis Correct and Complete?

Patients who have a secondary psychosis due to a medical or neurological disease (e.g., psychosis due to Alzheimer’s dementia) may not respond well or not at all to antipsychotics. A poor treatment response should alert you to a missed medical or neurological diagnosis. Other psychiatric conditions may also hinder an optimal response if only treated with an antipsychotic (e.g., comorbid OCD or psychotic depression). The presence of catatonic symptoms may require significant treatment modifications (e.g., using benzodiazepines and clozapine or ECT), both to achieve symptomatic improvement and to prevent a transition to lethal catatonia [8]. The most common reason for a poor response, however, is comorbid drug use. While a substance-induced psychotic disorder responds to antipsychotics, you rarely get a full response in the face of ongoing use, particularly if substance use is accompanied by poor adherence [9].

Has the Patient Had an Adequate Antipsychotic Trial?

Assuming you are confident that your diagnosis is correct and that drug use is not to blame for continuing symptoms, review past treatment trials with regard to adequacy. Table 12.1 summarizes the consensus definition of what constitutes an “adequate trial.” In reality, you are often left with some guess work regarding adequacy of past trials. Rarely do you have good documentation of dose, duration, and response; in almost all cases (unless the trial included a long-acting injectable antipsychotic), you should question adherence. A patient might report, “I was on risperidone and it did not work,” without further knowledge of dose, duration, adherence, and response.

First, the aforementioned TRRIP Working Group decreed that a good trial duration ought to be at least 6 weeks, which seems reasonable. You should expect some clearly detectable clinical response after 2–4 weeks of treatment (not counting a titration). In first-episode patients, however, time to response varies [10], with a small group of patients responding only after 8–16 weeks [11]. If there is a clear response, simply wait! It is important to allow symptoms to resolve and not to switch prematurely: time on (the right) medication is as important as the dose of medication since psychosis needs to be given time to resolve which will take several weeks. Similarly, functional benefits of adequate symptom control will accrue only with time; here we are measuring progress in months, not weeks. Giving more medication cannot accelerate this process. It is a mistake to “push the dose” after 1 week of hospitalizations, with little change in core psychotic symptoms. In other words, optimizing a medication can mean decreasing the dose for tolerability or adding something symptomatically or to counteract side effects. Most patients will in fact show signs of improvement as early as week 1 (or even after a single antipsychotic dose) (e.g., in sleep, in agitation) [12]. It is simply unrealistic to expect the resolution of a delusional system during a brief hospitalization.

Second, a dose of at least 600 CPZ-EQ should be used according to the Working Group. Some patients, however, might require a higher antipsychotic dose because of drug interactions (e.g., a patient on carbamazepine) or idiosyncrasies in drug metabolism, so do not just rely on the prescribed dose: therapeutic drug monitoring (TDM) is a critical aspect of managing patients who are considered treatment refractory (see Chap. 20 for TDM).

A higher than usual antipsychotic dose (higher than FDA-approved maximum dose) does not usually lead to further meaningful improvement. While clinicians often deploy a high-dose strategy, the only antipsychotics where clinical studies have shown a possible benefit is olanzapine [14]. Before giving up on olanzapine, consider therefore an olanzapine trial at a dose of 30 mg/day or more, particularly if clozapine is not an option. Note that this dose is higher than the Food and Drug Administration (FDA)-approved dose of 20 mg/day. An additional dose increase beyond 40 mg/d is unlikely to add more benefit in refractory patients [15]. With some antipsychotics, it has been confirmed in clinical trials that a higher dose is in fact not more effective than standard doses. Examples include ziprasidone (up to 320 mg/d) [16] and quetiapine (up to 1200 mg/d) [17, 18]. Keep in mind potentially dangerous problems when you exceed usual doses (e.g., dose-related QTc prolongation with quetiapine or ziprasidone).

Third, the Working Group proposed to use a cutoff of 80% adherence in order to be considered adequately adherent. If there is any chance that partial adherence might be the problem (I believe this is a very realistic possibility in almost all patients) and for a definitive conclusion of true treatment refractoriness, you may need to conduct a trial with a long-acting antipsychotic. This approach assures 100% adherence and, when coupled with TDM, confirms treatment-resistant schizophrenia if there is no response. I do not believe that oral trials with outpatients are sufficient.

Last, most guidelines, including the TRRIP Working Group guideline, recommend two well-conducted trials of an antipsychotic before moving to clozapine for refractory patients. I like to have one of the trials to be an antipsychotic with “tight” D2-binding like risperidone since you want to establish resistance to dopamine blockade. I hesitate to count quetiapine because of its transient and weak dopamine-2 (D2)-binding which could render it less effective for some patients toward a failed drug trial. If the two antipsychotics have not included olanzapine, I will often try olanzapine as it seems to be more effective than other first-line treatments [19]. The clinical task of establishing treatment resistance might not call for therapeutic adventures with newer antipsychotics where less is known clinically.

Important new information regarding the number of trials required in order to establish treatment resistance comes from a European trial called OPTiMiSE (which stands for Optimization of Treatment and Management of Schizophrenia in Europe) [20]. In OPTiMiSE, first-episode patients were treated according to a three-step algorithm. In the first phase, everybody received 4 weeks of initial treatment with the tight D2-blocking second-generation antipsychotic amisulpride (not available in the United States). In a double-blind randomized second phase, participants received either 6 more weeks with amisulpride, or they were switched to olanzapine. In the third phase, nonresponders received clozapine. The second phase asked the important question if switching antipsychotics is a useful strategy. The trial confirmed what we already knew: that first-episode patients as a group show a good response to antipsychotics – 2/3 of participants remitted after the end of the 10-week treatment with amisulpride or amisulpride and olanzapine [21]. The trial further showed that there was no additional benefit from switching to olanzapine; the patient who stayed on amisulpride did just as well. Finally, nonresponders benefited from switching to clozapine, but the response was not as robust as responders. What this trial basically shows is that some first-episode patients are refractory from the get-go, and the best intervention for this subgroup is a clozapine trial once it has been established that they are refractory to a first-line antipsychotic. This result confirms work by a Canadian group who has used a similarly stepped approach in their first-episode clinic. Like OPTiMiSE, patients who did not respond to their first antipsychotic benefited little from a switch to a second antipsychotic but showed a robust response once clozapine was offered [22]. Moving to clozapine after only one antipsychotic trial, particularly if treatment resistance was clearly established, may be something to discuss with your patient.