Blood Pressure Drugs May Hold Clues to Preventing Bipolar Disorder, Genetic Study Suggests

Could a class of drugs that millions of people already take for high blood pressure hold untapped potential for treating one of psychiatry’s most difficult conditions? A new study published in BMC Psychiatry suggests that diuretics, the water pills long used to manage hypertension, may be associated with a reduced risk of bipolar disorder. Using a genetic technique that sidesteps many of the pitfalls of conventional observational research, a team of Chinese researchers has uncovered evidence linking diuretic use to lower bipolar risk and has pinpointed two molecular targets that may explain the connection. The findings are preliminary, but they add momentum to one of the most intriguing ideas in modern drug development: that medicines designed for one organ system might quietly influence the brain.

The research, led by Jin Li and Mei-Fang Xu, with corresponding author Pei-Sen Yao of Fujian Medical University, employed a method called Mendelian randomization. Rather than following patients over time and comparing those who take diuretics with those who do not, an approach vulnerable to confounding and reverse causation, the team exploited natural genetic variation. Because genetic variants are randomly allocated at conception, much like a randomized clinical trial, they can serve as proxies for drug exposure. If people who carry variants that mimic the effects of diuretics show lower rates of bipolar disorder, that strengthens the case for a genuine causal relationship rather than a statistical coincidence.

The technical machinery behind the study drew on two large-scale genome-wide association studies, or GWAS. For the exposure side, the researchers used data on prescription diuretic use from the UK Biobank, a biomedical database containing hundreds of thousands of participants, together with blood pressure summary statistics from the International Consortium of Blood Pressure Genome Wide Association Studies. For the outcome side, they turned to bipolar disorder genetic data from the Psychiatric Genomics Consortium. Single-nucleotide polymorphisms, the individual letter changes in DNA that associate with diuretic use at the genome-wide significance threshold of P less than 5 times 10 to the minus 8, were selected as instrumental variables. These were then pruned for linkage disequilibrium using a stringent r-squared threshold below 0.001 within a 10,000-kilobase window, ensuring that each genetic instrument represented an independent signal rather than a duplicate copy of the same inherited segment.

The headline result was striking in its simplicity. Across the instrumental variable analysis, diuretic use showed a negative association with bipolar disorder risk, with an odds ratio of 0.93 and a 95 percent confidence interval of 0.88 to 0.98, reaching a P value of 6.39 times 10 to the minus 3. In plain terms, genetically proxied diuretic exposure was linked to roughly a 7 percent lower odds of bipolar disorder per unit of exposure. The estimate was calculated using inverse-variance weighting, the workhorse method of two-sample Mendelian randomization, which combines the effects of multiple independent genetic variants into a single causal estimate weighted by their precision. While a 7 percent reduction may sound modest, in the context of a condition that affects tens of millions of people worldwide and often resists existing medications, even small causal signals can point toward meaningful therapeutic avenues.

Crucially, the researchers did not stop at the association. They performed a battery of sensitivity analyses designed to test whether the signal could survive scrutiny. Horizontal pleiotropy, the phenomenon in which genetic variants influence the outcome through pathways unrelated to the exposure of interest, is the chief threat to the validity of any Mendelian randomization study. The team found no evidence of such pleiotropy for the identified associations. They also checked instrument strength using the F statistic, requiring values above 10 to rule out weak-instrument bias, a subtle distortion that arises when genetic proxies explain too little of the variation in the exposure. The genetic instruments passed these quality gates, lending credibility to the core finding.

To move from drug exposure to biological mechanism, the investigators turned to a complementary technique known as summary-data-based Mendelian randomization, or SMR. This method integrates expression quantitative trait loci, the genetic variants that regulate how strongly a gene is expressed, with GWAS data to ask whether a gene’s activity level itself influences disease risk. The analysis flagged two genes connected to diuretic-related biology. The first was ADO, a gene encoding adenosine deaminase, an enzyme involved in purine metabolism and adenosine signaling, a pathway with well-documented roles in neuromodulation. Lower expression of ADO was associated with a higher risk of bipolar disorder, with an odds ratio of 0.72, a 95 percent confidence interval of 0.60 to 0.87, and a P value of 4.71 times 10 to the minus 4.

The colocalization analysis provided the study’s most statistically robust moment. Colocalization asks whether the same causal variant drives both a gene’s expression and the disease association in a given genomic region, or whether two different variants merely happen to sit nearby. The team calculated a posterior probability of hypothesis 4, the scenario in which both traits share a single causal variant, at 92.4 percent for ADO, comfortably above the conventional 0.80 threshold for strong evidence. In other words, the genetic architecture of ADO expression and bipolar disorder risk appear to be genuinely intertwined, making ADO a credible molecular link between diuretic action and psychiatric vulnerability.

The second gene told a more cautionary tale. PRKCA, which encodes protein kinase C alpha, an enzyme central to intracellular signaling and previously implicated in mood regulation, showed the opposite direction of effect: higher expression was linked to greater bipolar risk, with an odds ratio of 1.21, a 95 percent confidence interval of 1.07 to 1.38, and a P value of 3.27 times 10 to the minus 3. However, the posterior probability of a shared causal variant between PRKCA expression and bipolar disorder was only 6.36 percent, far below the threshold for confidence. The association with PRKCA therefore remains suggestive rather than established, a reminder that even statistically significant findings in genetic epidemiology can dissolve under closer inspection when linkage cannot be distinguished from causation.

The authors are careful to frame their conclusions as hypothesis-generating rather than practice-changing. In their own words, the findings identify diuretic-related molecular targets associated with bipolar disorder risk and provide potential hypotheses for further investigation of the biological mechanisms and therapeutic relevance of these targets, but they should be considered target-discovery evidence that requires further functional, pharmacological, and clinical validation. This caution is well placed. Mendelian randomization estimates the effect of lifelong, genetically influenced exposure, which may not correspond to the effect of starting a diuretic in midlife. Moreover, the study relied on summary-level data from previously published GWAS and did not involve direct participation of human subjects, meaning no individual patient records were examined. Clinical trials would be needed before any diuretic could be considered for psychiatric use.

Nevertheless, the study exemplifies a growing strategy in drug repurposing, the search for new uses for old medicines. Repurposed drugs arrive with known safety profiles, established manufacturing, and in some cases decades of real-world evidence, dramatically shortening the path from discovery to patients. The work, supported by the Fujian Provincial Science and Technology Innovation Joint Fund and conducted under UK Biobank application number 94695, also highlights the power of publicly available genetic datasets to generate psychiatric leads from the comfort of the computational laboratory. Whether diuretics ultimately earn a place in the bipolar disorder treatment arsenal remains an open question, but the study has supplied something psychiatry desperately needs: concrete, genetically grounded molecular targets, chief among them ADO, around which future experiments, drug screens, and perhaps one day clinical trials can be built.

Subject of Research: Genetic evidence for repurposing diuretics as a potential therapy for bipolar disorder

Article Title: Repurposing diuretics for bipolar disorder: a Mendelian randomization study

Article References: Li, J., Xu, M.-F., Ye, R.-T., Wang, S.-H., & Yao, P.-S. (2026). Repurposing diuretics for bipolar disorder: a Mendelian randomization study. BMC Psychiatry. https://doi.org/10.1186/s12888-026-08581-1

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08581-1

Keywords: bipolar disorder, diuretics, Mendelian randomization, drug repurposing, genetics, GWAS, UK Biobank, ADO gene, PRKCA, Psychiatry, colocalization, molecular targets