Differential genetic correlations across major psychiatric disorders between Eastern and Western countries.

Abstract

Recent findings from genetic correlation analyses clarify the complicated relationships regarding shared genetic risk among psychiatric disorders. Using the datasets of the Psychiatric Genomics Consortium [PGC; mainly consisting of European (EUR) samples],1, 2 genetic correlations (rg) among bipolar disorder (BD) subtypes I/II (BD1/BD2), schizophrenia (SCZ) and major depressive disorder (MDD) revealed (1) the highest rg between BD1 and BD2 (rg = 0.85), (2) a higher rg between BD1 and SCZ (rg = 0.66) than that between BD2 and SCZ (rg = 0.54) and (3) a lower rg between BD1 and MDD (rg = 0.34) than that between BD2 and MDD (rg = 0.66).1 To replicate these findings, we aimed to extend the genetic relationships between BD subtypes and SCZ or MDD in East Asians (EAS): we calculated rg based on Linkage Disequilibrium score regression analysis3 using genome-wide association study data for BD1/2,4 SCZ,5 and MDD6 from EAS (Japan and China: Supplementary Text). To compare the trends in EAS, we referred to the EUR results by PGC.1, 2 We found a high genetic correlation between BD1 and BD2 in EAS, with a lower magnitude than the finding from PGC.1 However, different relationships were observed in EAS between BD1 and SCZ/MDD and between BD2 and SCZ/MDD (Figure 1 /Table S1). Specifically, we observed (1) a higher rg between BD1 and MDD in EAS (‘mood’ spectrum), compared with the PGC results,1 where BD1 was closer to SCZ (“psychosis” spectrum), (2) a lower rg between BD1 and SCZ in EAS, but (3) confirmed similar trends of rg between BD2 and SCZ/MDD. If the genetic risk of psychiatric disorders follows the common disease common variant (CD-CV) hypothesis, the genetic components are shared across populations; however, we detected a discrepancy in genetic correlations between BD1/BD2 and SCZ/MDD among different ethnic populations. The current results may correspond to the discrepant trend in the prevalence of BD and SCZ. Previous epidemiological surveys reported a much lower prevalence of BD1 in EAS (~0.1%)7 compared with that in EUR (~1%),7 although BD has often been underdiagnosed and/or underreported. However, unlike BD, the prevalence of SCZ has been reported at approximately 1% worldwide, regardless of EAS and EUR. According to our results, as well as to the CD-CV hypothesis and epidemiological evidence, we reconsidered the diagnostic attitude towards BD1, specifically the boundaries between ‘mania with psychotic features’ (restricted to BD1 diagnosis) and other psychotic disorders; we speculate that the BD1 diagnosis in Japan adheres to ‘mood’ symptoms (e.g., mania), conversely the one in EUR is more liberal regarding ‘psychosis’ in manic phase. In this case, especially for BD1, Japanese psychiatrists may prefer to diagnose patients (and recruit them into the ‘study’) with ‘extreme’ manic symptoms ‘without’ psychotic features. Further, we assume this stance may be associated with the high genetic correlation (rg = 0.54) between BD1 and MDD in EAS. The different tendency in EUR resulted in the high rg between BD1 and SCZ in the PGC (rg = 0.66), since SCZ liability was enriched in BD with psychotic features (mood-incongruent > mood-congruent).8 To clarify, we surveyed the ratio of BD1 with psychotic features and found that our sample contained only approximately 30% of BD1 patients diagnosed with ‘with psychotic features’ (i.e., 177 ‘BD1 with psychotic features’ out of 631 accessible BD1 samples). In contrast, a previous study reported that half of BD1 subjects (among them approximately 2/3 ~ mood-congruent, 1/3 ~ mood-incongruent) in EUR exhibited psychotic features,9 resulting in a larger difference in the ‘psychotic features’ ratio between EAS and EUR compared to the difference of BD1 prevalence (the estimated prevalence of ‘BD1 with psychotic features’ in Japan: 0.1% × 0.3≂0.03%, that in EUR: 1% × 0.5≂0.5%). Of course, these differences may be explained by a ‘simple’ sampling bias of our datasets; however, we must be aware of the possibility that ‘diverse’ ascertainment bias (due to different diagnostic trends in different cultures) may have influenced the sample constitution in the genetic study and, moreover, possibly in clinical studies. Nevertheless, it must be emphasized that either attitude (towards ‘mood’ or ‘psychosis’) is not always wrong; the BD1 datasets from Japan (and possibly EAS) comprise samples diagnosed with emphasis on ‘mood’ symptoms, which is appropriate for the aim of ‘dissecting’ BD1 vs. SCZ (typical ‘psychosis’). Conversely, the diagnosis of BD1 in Western countries potentially achieves capturing a wide range of BD patients, because they ‘straightforwardly’ followed the DSM/ICD criteria. Therefore, these results suggest that symptomatology-based dissection is key to define novel diagnostic architecture that could standardize the diagnosis of BD1/BD2 and SCZ.10 In conclusion, we detected different genetic correlations among psychiatric disorders across populations. The diagnostic attitude may explain these findings, although further analysis is required to evaluate other possibilities. The authors thank members of the Psychopathology group at the Department of Psychiatry, Nagoya University Graduate School of Medicine for their helpful and critical comments. They have no conflicts of interest to declare. The authors also thank MARUZEN-YUSHODO Co., Ltd. for the English language editing. This study was supported by the SRPBS from the Japan Agency for Medical Research and Development (AMED) under Grant Numbers JP20dm0107097 (NI, MI, CT and TS), JP22wm0425008 (NI, MI and TS) and JP22wm0525024 (MI and TKishi); GRIFIN of P3GM from AMED under Grant Numbers JP20km0405201 (TS and NI), JP20km0405208 (MI), JP22tm0424220 (CT and MI); AMED under Grant Number JP22dk0307107 (TKishi); Health and Labor Sciences Research Grant under Grant Numbers 20GC1017 (TKanazawa) and 21GC1018 (TKishi); JSPS Kakenhi Grant Numbers JP21H02854 (MI), JP19K08082 (TKishi), JP18K15497 (TS), and JP21K07490 (TS); and the Private University Research Branding Project from MEXT (NI). The funders played no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Dr. T Kanazawa received research support from Eisai, Sumitomo and Otsuka. Dr. T Kato has received research support or speaker's honoraria from the Japan Agency for Medical Research and Development (AMED), MEXT, Sumitomo Pharma, Otsuka Pharmaceutical, Takeda Pharmaceutical Co., Ltd., Eisai Co., Ltd. and Teijin Pharma; consulting fees from Sumitomo Pharma and Janssen Pharmaceutical K.K.; speaker's honoraria from Sumitomo Pharma, Otsuka Pharmaceutical, Taisho Pharma Co., Ltd., Meiji Seika Pharma, Mochida Pharmaceutical Co., Ltd., Shionogi & Co., Ltd., Janssen Pharmaceutical K.K., Janssen Asia Pacific, Yoshitomiyakuhin, MSD K.K., Kyowa Pharmaceutical Industry Co., Ltd., Takeda Pharmaceutical Co., Ltd., Taisho Pharmaceutical Co., Ltd., Eisai Co., Ltd., Viatris and Mylan N.V. and is in leadership positions in Psychiatry and Clinical Neurosciences (Editor in Chief), Brain Science Working Group of MEXT (Chair), Japanese Society of Biological Psychiatry (president), Bipolar Disorder Committee, Japanese Society for Mood Disorders (Chair) and Japanese Society for Neuroscience (Vice President). Dr. T Kishi has received speaker's honoraria from Sumitomo, Otsuka, Takeda, Eisai, Janssen and Meiji. Dr. N Iwata received research support or speakers' honoraria from Sumitomo, Eisai, Daiichi Sankyo, Takeda, Meiji, Tanabe-Mitsubishi, Otsuka, Eli Lilly, Janssen and Viatris. The remaining authors declare no conflicts of interest. Table S1 Genetic correlations between subtypes of bipolar disorder (BD1/BD2) and schizophrenia/major depressive disorder (SCZ/MDD) in East Asian and European populations. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.