The authors used summary-data-based Mendelian randomization to model genetically predicted pharmacological actions of topiramate. They derived cis-expression quantitative trait loci (cis-eQTLs) from brain and blood tissue to create instruments for genes encoding the drug’s molecular targets. Five instruments met the validation criteria the study team applied: brain expression of CA4 and CACNB3, and blood expression of CA4, GRIK1, and GRIK3. These instruments were presented as consistent with topiramate’s antiepileptic action, according to the manuscript.
Short, focused point: instruments came from both brain- and blood-derived eQTL data, not from new clinical samples. The source reported that all data were publicly available summary statistics.
Data indicated that genetically predicted inhibition of CA4 expression in blood associated with decreased risk of binge-eating behaviour (BE). The authors reported a statistically significant effect after multiple-testing correction (p = 0.002) and found that the association remained significant using SMR-multi (p = 0.0049). Colocalization analysis provided moderate support for a shared causal variant (reported PPH4 = 0.58) linking CA4 expression and BE. The study team treated this signal as the most robust positive finding from their analyses.
This result was specific to blood-derived CA4 expression in the authors’ analyses. The manuscript did not claim a demonstrated clinical effect of pharmacological topiramate exposure on BED diagnoses in treated patients; rather, investigators interpreted the genetic evidence as motivating further clinical and functional work to clarify mechanisms of appetite suppression and weight loss associated with the drug.
The authors tested primary outcomes of anorexia nervosa (AN; 24,223 cases and 1,243,971 controls) and BE (39,279 cases and 1,227,436 controls), and they added eight exploratory psychiatric disorders to the analysis set: attention-deficit/hyperactivity disorder, anxiety disorders, alcohol use disorder, bipolar disorder, major depressive disorder, panic disorder, post-traumatic stress disorder, and schizophrenia. Following multiple-testing correction and the suite of robustness checks, the manuscript reported no robust evidence that genetically predicted perturbation of the tested drug targets influenced AN risk. Likewise, the exploratory psychiatric outcomes showed no robust associations in the authors’ analyses.
The authors emphasized the single, reproducible signal for CA4 in blood as distinct from the null results for AN and the other psychiatric phenotypes. They did not report additional positive target–outcome links that survived correction and sensitivity analyses.
Investigators applied SMR to link gene expression (from cis-eQTLs) with case–control summary statistics for the psychiatric outcomes. They validated five instruments consistent with topiramate target biology and assessed robustness using SMR-multi, HEIDI, and colocalization analyses. The source reported these methods as the primary means to evaluate instrument specificity and to probe shared causal variants between expression and outcome.
The manuscript noted that all data used were publicly available summary statistics and that the analyses did not involve new patient recruitment or primary phenotyping. The study team declared no competing interests and confirmed adherence to relevant ethical reporting guidelines given the secondary-data nature of the project.
The authors concluded that genetically predicted inhibition of CA4 in blood associated with lower risk of binge-eating behaviour, and that colocalization provided moderate support for a shared causal variant. They cautioned that no robust effects were found for anorexia nervosa or the eight exploratory psychiatric disorders they assessed. The manuscript explicitly framed results as hypothesis-generating: investigators recommended further clinical and functional investigations to clarify mechanisms underlying topiramate-associated appetite suppression and weight loss and to determine relevance to eating-disorder risk.
Key caveats documented in the source: this work is a preprint and has not been certified by peer review; the analyses used publicly available summary statistics rather than randomized or interventional exposure data; and the authors did not report any clinical dosing, temporal exposure, or direct pharmacokinetic–pharmacodynamic measurements linking genetic proxies to drug effect sizes. The study team therefore advised against using these findings to guide clinical practice without additional validation.
Unanswered questions left by the manuscript include which molecular or cellular mechanisms would causally link CA4 modulation to behavioural changes in eating, whether blood-derived eQTL signals translate to actionable pharmacology in target tissues, and what clinical trial evidence—if any—could directly test topiramate’s effect on eating-disorder incidence or symptom trajectories. The authors recommended follow-up clinical and functional studies to address these open questions.