A SYSTEMATIC REVIEW AND META-ANALYSIS OF CLINICAL VARIABLES ASSOCIATED WITH RESPONSE TO CLOZAPINE IN TREATMENT RESISTANT SCHIZOPHRENIA

Abstract Background Approximately one third of patients with schizophrenia display suboptimal response to two trials of non-clozapine antipsychotic medication and may be termed treatment resistant. Clozapine is the only licensed pharmacotherapy for treatment resistant schizophrenia, but response to clozapine is variable and can only be determined through a trial of treatment. Understanding demographic and clinical sources of varied response to clozapine may be useful in the optimisation of clinical treatment algorithms and stratification of patient groups for clinical trials of early use of clozapine. Methods We systematically reviewed literature to investigate clinical and demographic factors associated with variation in clozapine response. Articles were eligible for review if they reported differences in clozapine response as a function of baseline variables within patient samples of schizophrenia spectrum disorders. In a second step, a random-effects meta-analysis to study group mean differences in age, age of onset and duration of illness between clozapine responders and non-responders was performed. Results Thirty-one articles were eligible for qualitative review. The systematic review found that a poorer response to clozapine was associated with older age at clozapine initiation, younger age at illness onset and longer delay in clozapine initiation. A higher number of previous hospitalisations and antipsychotic trials prior to treatment with clozapine were also associated with poorer outcomes. Both systematic review and meta-analysis identified that longer durations of illness before clozapine initiation were associated with worse clinical outcomes. In a total sample of 313 participants (n = 158 responders), clozapine responders had a significantly shorter duration of illness than non-responders (g = - 0.31; 95% CI, 0.06 - 0.56; p = 0.02). Analysis was then limited to studies with a minimum follow-up period of 12 weeks to align with the recommended amount of time to monitor clozapine response. The difference in duration of illness between responder versus non-responder groups remained significant and overall effect size increased (g = - 0.42; 95% CI, 0.17 – 0.67; p < 0.001). Between study heterogeneity was low (Q = 3.59, I2 = 0%, P = 0.61). Discussion The results imply that delay in clozapine treatment is associated with worse response and support the view that initiation of clozapine earlier in illness may be beneficial. Although we cannot make causal assertions from the data presented, poor response to clozapine when prescribed after a longer delay is likely due to a combination of factors; including the effects of sustained active symptoms on neurobiological integrity, social functioning and self-care. Given evidence that early non-response to conventional antipsychotics predicts a later diagnosis of treatment resistance and poorer outcomes, identifying treatment resistance and prescribing clozapine earlier in the illness course would prevent unnecessary loss of time in the treatment of refractory psychosis. Current research into clinical variables associated with clozapine response is limited to few studies but future investigation into the predictive value of these variables is warranted. This is important given the relative ease and low-cost clinical information can be obtained from acutely unwell patient groups who may poorly tolerate more invasive research and clinical procedures.