The comorbidity between urinary incontinence (UI) and mental disorders which included attention deficit (ADHD), bipolar disorder (BIP), major depressive disorder (MDD), post-traumatic stress disorder (PTSD) and schizophrenia (SCZ) remains unclear in observational studies. We aimed to clarify this association using a large-scale genome-wide cross-trait analysis.

Leveraging summary statistics of large-scale genome-wide associations studies (GWAS) conducted in European-ancestry populations on UI (N=393786), ADHD (N=55374), BIP (51710), MDD (N=172705), PTSD (N=174659), SCZ (N=135236). We first applied genetic correlation analysis using linkage disequilibrium score regression(LDSC), followed by the multi-trait analysis of GWAS (MTAG) to identify the trait-specific SNPs. We then performed SNP-level functional annotation to identify significant genomic risk loci. Transcriptome-wide association analysis (TWAS) was implemented to explore novel UI-associated genes, followed by pathway enrichment analysis to understand underlying biological mechanisms, and we used the BH correction for each trait’s all gene-tissue pairs on TWAS P values to account for multiple testing. Furthermore, we conducted bidirectional Mendelian randomization (MR) and latent causal variable (LCV) model to test the potential causal relationships between each mental disorder and UI.

The LDSC results found that UI shared positive correlations with ADHD(Rg=0.4454, P=4.34E-14), BIP (Rg= 0.1718, P= 1.81E-03), MDD (Rg= 0.357, P= 8.04E-11), PTSD (Rg= 0.4077, P= 3.23E-05) and negative correlation with SCZ (Rg= 0.065, P= 1.10E-01). 
5 independent SNPs shared between UI and ADHD were identified by MTAG. The most significant SNP (rs8034190, PUI=3.00E-05, PADHD= 9.22E-05, PMTAG= 1.25E-07) was located at the ncRNA-intronic region, which was near gene SEMA6D. 11 independent SNPs shared between UI and BIP were found. The most significant SNP (rs6088520, PUI=7.40E-07, PADHD= 3.87E-03, PMTAG= 7.39E-09) was located at an intergenic region and the second significant SNP (rs12474906) was located at the intronic area and mapped to MRPL33:RBKS, which was also shared between UI and SCZ ((sentinel SNP: rs13001060, PUI= 2.70E-05, PSCZ= 2.43E-06, PMTAG= 2.06E-09). We identified 11 independent SNPs shared between UI and MDD, and the most significant SNP was rs1554825 (PUI= 1.30E-03, PMDD= 8.70E-06, PMTAG= 3.15E-07), located at a ncRNA_intronic region, whose mapped gene was TEX41. 16 independent SNPs shared between UI and PTSD and the most significant SNP (rs74864822, PUI=4.70E-05, PPTSD= 2.49E-05, PMTAG= 1.84E-08) was located near gene LINC01470. For UI and SCZ, 19 shared independent SNPs were identified, among these the most significant SNP (rs12774693, PUI= 3.70E-03, PSCZ= 3.44E-13, PMTAG= 9.44E-10) was located at an intronic and mapped to SUFU gene. As shown in Table 1.
The GTEx enrichment analysis independent tissue expression that was significantly enriched (after Bonferroni correction) for expression of cross-trait associated genes for SCZ and UI, including blood, heart, brain, and muscle. In GO terms, we found there were four pathways that UI shared with five mental diseases, which were immune system development, regulation of cell differentiation, regulation of immune system process, and side of the membrane, respectively, implicating that processes involving immunity may account for shared causes.
We identified 8 and 1 TWAS-significant genes between UI and BIP and MDD, respectively. Specifically, the PGAP3 gene, located at 17q12, was shared by UI (PTWAS.BH=0.006) and BIP trait (PTWAS.BH= 0.042), and the gene was also found to be genome-wide significant in UI and BIP cross-trait meta-analysis (PMTAG=4.40E-07). MAP1LC3A was shared by UI(PTWAS.BH=0.041) and BIP(PTWAS.BH=0.036), and it was genome-wide significant in UI and BIP cross-trait meta-analysis (PMTAG=7.39E-09). CCDC152, a TWAS-significant gene between UI (PTWAS.BH=0.006) and MDD(PTWAS.BH=0.015), located in the antisense region of the SELENOP gene which encodes a major plasma selenoprotein. It was also genome-wide significant in UI and MDD cross-trait meta-analysis (PMTAG=5.41E-07). 
LCV model analysis showed that no trait pairings exhibited strong evidence for a putative causal relationship between ADHD and UI (GCP estimate = 0.038, SE = 0.553, P = 0.834), BIP and UI (GCP estimate = -0.012, SE = 0.257, P = 0.556), SCZ and UI (GCP estimate = 0.151, SE = 0.547, P = 0.562). which means there may be a bidirectional relationship among these trait pairings. For MDD and UI (GCP estimate = 0.416, SE = 0.345, P = 0.394), the low GCP estimate which didn’t reach the intermediate GCP estimate showed that they have limited partial genetic causality. For PTSD and UI (GCP estimate = 0.565, SE = 0.259, P = 0.047), the GCP marginally failed to exceed the mean posterior GCP ≥ 0.6 threshold but exceeded the intermediate GCP = 0.43 threshold. Our IVW results showed that liability to mental disorders does not affect UI risk, though the contamination mixture results showed significant evidence of the causal effect of BIP on UI (ORconmix=1.090, 95%CI 1.016-1.229, Pconmix=0.001), this result was not consistent with the results from other MR methods. Furthermore, we performed the reverse MR analyses with liability to UI as the exposure and risk of 5 mental diseases as outcomes and we found no significant association.

We conducted large-scale genome-wide cross-trait analyses that investigates the shared genetic basis underlying UI and mental disorders. In genetic correlation analysis, we found positive genetic correlations between UI and ADHD, BIP, MDD, and PTSD and a negative correlation with SCZ. We also identified the genetic overlap between UI and mental disorders at the individual variants level, including 5 loci shared by UI and ADHD, 11 loci shared by UI and BIP, 11 loci shared by UI and MDD, 16 loci shared by UI and PTSD and 19 loci shared by UI and SCZ. Furthermore, we investigated whether shared genes between UI and mental disorders have a potential functional connection with the human tissues. In the TWAS analysis, we found multiple shared tissue-gene pairs UI and BIP and MDD, respectively, including the exocrine, nervous, digestive, respiratory and immune systems. Besides, we investigated causal relationships between UI and mental disorders using MR and LCV. Our LCV results suggested there may be a bidirectional relationship among ADHD-UI, BIP-UI and SCZ-UI trait pairings, while MDD and PTSD might increase the risk of UI, providing insights into the pathological mechanisms of UI.

Our findings suggest shared genetics and causation between UI and mental disorders, and advance our understanding of the genetic overlap between them. Identifying shared drug targets and molecular pathways can be beneficial to treat UI and mental disorders more efficiently.

Continence 7S1 (2023) 100880
DOI: 10.1016/j.cont.2023.100880