Long-Term Real-World Effectiveness of Pharmacotherapies for Schizoaffective Disorder

Jonne Lintunen; Heidi Taipale; Antti Tanskanen; Ellenor Mittendorfer-Rutz; Jari Tiihonen; Markku Lähteenvuo


Schizophr Bull. 2021;47(4):1099-1107. 

In This Article


It is of great clinical importance to study purely SCHAFF patients, as a plethora of treatment modalities are used in clinical practice, although only paliperidone has an approved treatment indication for SCHAFF from EMA and FDA, and use of other treatments would be considered off-label use. As might be expected, exposure to antipsychotics, in general, was associated with a decreased risk of psychosis hospitalization. Concerning specific antipsychotics, exposure to LAIs or clozapine was associated with a lower risk than FG- and SG-oral pharmacotherapies in both nationwide cohorts. In addition, especially in the Swedish cohort, antipsychotic polytherapy was associated with a notable decrease in the risk of psychosis hospitalization. Of specific antipsychotics, use of levomepromazine or aripiprazole LAI was associated with a decreased risk of psychosis hospitalization in the Finnish cohort, but not in the Swedish cohort. For aripiprazole LAI, this could be due to the low number of users and as such, this result should be interpreted with caution. The risk of treatment failure (psychiatric hospitalization, any change in medication, death) was increased in the Swedish cohort during levomepromazine use, but it had no effect on the risk of treatment failure in the Finnish cohort. This could be due to levomepromazine's poor antipsychotic effect[20] and the possibly increased risk of mortality during levomepromazine use.[21]

The combination of antipsychotics and adjunctive mood stabilizers was associated with a lower risk of psychosis hospitalization in both cohorts as compared with antipsychotic monotherapy, which is important to note, as combining medications from two different groups can increase the risks associated with the treatments[22] and reduce adherence, and should thus not be undertaken without evidence on effectiveness.

The superiority of LAIs and clozapine is in line with previous studies that compared different antipsychotic treatments in large cohorts combining both schizophrenia and SCHAFF patients.[16,17,23] The superior effectiveness of LAIs[24,25] and clozapine[26] for schizophrenia has also been observed in meta-analyses. The effectiveness of these drugs is, at least to some extent, likely associated with better treatment adherence. This is in line with our observation on the risk of treatment failure; the risk associated with LAIs and clozapine was lower as compared with oral pharmacotherapies. Nonadherence to antipsychotic medications has been estimated to be between 41% and 50%.[27] However, adherence to LAIs[28] and clozapine[29] has been shown to be better than adherence to other antipsychotics, which could be due to regular appointments with healthcare professionals for administering antipsychotic injections and obligatory blood monitoring during clozapine use. Therefore, possible exacerbations of SCHAFF are likely more rapidly noticed and outpatient interventions are undertaken to prevent inpatient psychiatric hospitalization. However, our results show some differences between the effectiveness of different LAIs, and therefore, the superiority of LAIs may not be explained solely by better adherence and more frequent outpatient healthcare visits. In addition to clozapine's effectiveness on treatment-refractory psychotic symptoms, exposure to clozapine has been associated with a decreased risk of suicidal behavior and substance abuse, and it has been suggested to be especially effective on affective symptoms.[30,31] All of these factors relate to better treatment outcomes in SCHAFF.

Since SCHAFF symptomatology includes both schizophrenic and affective symptoms, adjunctive mood stabilizer or antidepressant use is common, even though the evidence-based effectiveness of this practice is inconclusive.[9–11] It has been proposed that the optimal treatment for the different subtypes of SCHAFF might be different; for the manic/bipolar subtype use of mood stabilizers could be useful, whereas the depressive subtype could benefit from antidepressants.[3] Previous studies on the efficacy of adjunctive use of mood stabilizers in SCHAFF have been small and they have focused on specific combinations of antipsychotics and mood stabilizers.[32–35] Indeed, preceding information on the use of mood stabilizers as adjunctive therapy in SCHAFF has been scarce and due to limitations, very few conclusions could have been drawn. Our results suggest that add-on mood stabilizer use could be beneficial in SCHAFF. However, our study did not separate the different subtypes of SCHAFF, which could have yielded more clinically relevant information.

Evidence on adding antidepressants to schizophrenia/SCHAFF pharmacotherapy is also limited.[36] Our results differed between the two cohorts, as in the Swedish cohort exposure to antidepressants was associated with a decreased risk of psychosis hospitalization, whereas in the Finnish cohort this association was not observed. These differences could be due to different treatment guidelines and clinical practices between Finland and Sweden. It is also worth noticing that using BZDRs with antipsychotics was associated with an increased risk of psychosis hospitalization in both cohorts. This is especially important, as BZDRs were very widely used in both of our study cohorts. It is concerning that BZDRs were more commonly used than other adjunctive medications, even though the results suggest beneficial effects of mood stabilizers and antidepressants.

The strengths of this observational study include two large nationwide cohorts with thousands of patients and multiple years of follow-up time. Selection bias is minimal, since our study included all patients with SCHAFF in the nationwide registers (excluding a small number of patients who have only been treated in outpatient care in Finland). Therefore, our results represent a real-life setting and are generalizable to high-income countries that provide medications for free or with very low copayment for patients with serious mental disorders. Drug use was modeled with the PRE2DUP method[18] which has been shown to produce highly reliable estimates of drug use.[37] The within-individual model used in this study controls for all time-invariant covariates in the design, and the analyses were adjusted for multiple time-varying covariates, minimizing the common sources of bias in observational studies. The limitations of our study are related to the nature of the data in the nationwide registers, which were not originally designed for analysis methods such as the ones employed here. Register-based data lack information on many clinically important factors, such as the severity of the symptoms during specific drug exposures and thus, residual confounding may exist. We were not able to confirm the certainty of the diagnostics between SCHAFF and schizophrenia from the medical records since our study is register-based. However, SCHAFF- and schizophrenia diagnoses are carefully considered before the diagnosis is given, and in both Finland and Sweden diagnostics should always follow the ICD criteria. Patients with diagnostic uncertainties (change of diagnosis from SCHAFF to schizophrenia) were removed from the sensitivity analyses. One weakness of the current study was that we did not make a division according to the subtypes of SCHAFF, as the different subtypes or polarities of episodes (manic, depressive, or mixed) may respond differently to pharmacological treatments. This topic remains for future studies to elucidate upon. Psychological interventions are important when treating psychiatric disorders, but unfortunately, our data did not allow us to adjust our models for them. It is also noteworthy that both Finland and Sweden use the ICD-system, and therefore, these results may not be directly translatable to healthcare systems using DSM criteria.

In conclusion, our study found that use of antipsychotics, especially clozapine and LAIs, was associated with a decreased risk of psychosis hospitalization among individuals with SCHAFF in two nationwide cohorts. These results are in line with previous nationwide cohort studies, which combined both schizophrenia and SCHAFF patients. Add-on mood stabilizer treatment was associated with a significant decrease in the risk of psychosis hospitalization as compared with antipsychotic monotherapy in both cohorts, whereas use of antidepressants had varying results. Use of benzodiazepines with antipsychotics was associated with an increased risk of psychosis hospitalization. SCHAFF is a common, yet controversial, diagnosis in clinical psychiatry, and there is a gap between research-based knowledge and clinical practices on its pharmacotherapy. Therefore, more studies including only patients with SCHAFF are needed to determine the optimal pharmacotherapy for this disorder. Especially studies on adjunctive treatments should be performed, and the subtypes of SCHAFF should be separated, since the treatment response could differ among different types of SCHAFF patients or between episodes of different polarities.