As expected, the present study showed high variability in OLZ rough plasma concentrations values (59.54 ± 35.80 ng/mL), probably due to the wide range of OLZ daily doses patients were taking (15.45 ± 5.25 mg/d). Dose-corrected concentration (C/D) was used to eliminate the interference of dose and observe the influence of other factors on drug concentration. These results are similar to those of previous studies(Fekete et al. 2017;Gex-Fabry et al. 2003). This study simultaneously revealed the influence of CYP1A2, CYP2D6, CYP3A5, CYP2C19, UGT2B15, UGT1A4, and UGT2B7 variants, as well as individual information on body weight, renal function, and co-treatment on C/DOLZ in patients with schizophrenia. The results illustrated that gender, co-treatment with VPA, and UGT1A4 genotypes significantly affected C/DOLZ in patients with schizophrenia.
OLZ is metabolized by uridine diphosphate-glucuronosyltransferases (UGT) and cytochrome P450 (CYP) enzymes(Soderberg and Dahl 2013;Haslemo et al. 2012b;Callaghan et al. 1999). The most important metabolic pathways for OLZ are UGT1A4 genotypes and UGT2B10 genotypes(Erickson-Ridout et al. 2011), significantly impacting OLZ glucuronidation. The other essential pathways are CYP enzymes. CYP1A2 plays a significant role in oxidation among all the CYP enzymes, and CYP2D6 has a minor contribution to oxidation(Korprasertthaworn et al. 2015). CYP1A2 genotypes appeared to have a significant influence on Olanzapine concentrations (Hattori et al. 2020;Djordjevic et al. 2020). Due to our naturalistic design, we include UGT2B7 and UGT2B15 genotypes, but unfortunately, UGT2B10 polymorphism was not measured. The one-way analysis of variance (ANOVA) found significant differences in C/DOLZ among UGT1A4, CYP1A2*1F, and CYP2D6*3 genotypes. UGT2B15 genotypes did not show their significance (P = 0.056) in ANOVA, but more considerations are still needed. A critical pathway of OLZ is N-glucuronides reactions, and UGTs catalyze many N-glucuronides. Finally, as independent variables, CYP1A2*1F, CYP2D6*3, UGT1A4, UGT2B7, and UGT2B15 variants were measured in multiple linear regression analysis. The results showed that the UGT1A4 genotypes significantly affected C/DOLZ. TT genotype carriers of UGT1A4 had a higher C/DOLZ than GT genotypes. The study found an association between UGT1A4 genotypes and OLZ dose-corrected concentration, which requires further verification by expanding the sample size.
We also found that COLZ and C/DOLZ were significantly higher in OLZ treated with VPA concomitantly in ANOVA, and the multiple linear regression analysis also showed its significance. It reminds us that co-treatment with VPA needs more attention in clinical deterioration in prescribing OLZ. The knowledge of the interacting mechanism of VPA-reducing COLZ is too limited. The plausible explanations might be as follows(Zhang et al. 2019;Tveito et al. 2019): VPA reduced COLZ by induced P-glycoprotein (P-gp) efflux in the intestine; VPA altered enterohepatic recirculation of OLZ; OLZ and VPA shared metabolic pathways such as UGT2B10. Patients with schizophrenia often need the combination therapy of OLZ and VPA. However, the changes in OLZ exposure might worsen the symptom because co-treatment showed lower plasma concentrations. Moreover, these two drugs' adverse drug reactions (such as sedation, weight gain, and tremor) partly overlap and lead to deterioration. As olanzapine plasma concentrations are highly variable, co-treatment with VPA and OLZ strengthens the instructions for TDM as a tool for dose adjustments of OLZ.
Thirdly, we found that the C/DOLZ was higher in females than in males. Generally, females have more diminutive size of livers and the kidneys could explain the difference. We also know that CYP1A2 has a lower activity in females, and CYP1A2 is the main enzyme in OLZ metabolism. The gender effect of OLZ indicated that gender should be considered when prescribing OLZ.
This study has some limitations. Firstly, this study did not consider the clinical outcome data and adverse drug reaction, so we cannot infer whether dosage and the variations on C/DOLZ impact OLZ response. Secondly, the small sample size resulted in a few samples with some genotypes, leading to lower statistical power. It would be appropriate to enrich the sample size to increase patient samples in low prevalence genotype carriers to improve the statistical power. Thirdly, our study did not measure the metabolites of OLZ, and we cannot calculate the capacity of OLZ metabolic enzymes directly. Future studies can simultaneously determine OLZ and its metabolites to explore the relationship between OLZ concentration and clinical outcomes. Fourthly, Due to the naturalistic design of our study, one of the critical metabolic enzymes, UGT2B10 genotypes, was not measured. The polymorphisms in the ABCB1 gene associated with P-glycoprotein function may also affect OLZ plasma concentrations, which were not measured in this study either. Fifthly, inflammation may increase the activity of CYP1A2, but we did not collect the effect of inflammation in this study.
In contrast, this study still showed several strengths. Firstly, all the enrolled patients were at quit smoking status for more than four weeks or none smoker. Secondly, all patients had the same diet conditions. This study controlled the effects of smoking and diet on OLZ concentrations.