Suicide is a major public health concern with increasing rates globally, particularly in the United States. Identifying causal risk factors is crucial for developing effective prevention strategies. Observational studies have linked lower educational attainment (EA) and cognitive performance (CP) to increased suicide attempt risk, but these relationships are confounded by various factors. This study aimed to address this limitation by using Mendelian randomization (MR), a genetic epidemiology method that leverages naturally occurring genetic variations as instrumental variables to infer causal relationships. MR is particularly useful when traditional randomized controlled trials (RCTs) are unethical or impossible to conduct. The study hypothesized that lower EA and/or lower CP might causally influence suicidal behavior. The authors note the complex interplay of genetic, psychological, social, and cultural factors contributing to suicidal behavior and the polygenic and heritable nature of both EA and CP. The study used the largest publicly available GWAS data on EA, CP, and suicide attempt risk to rigorously examine these relationships, employing both single-variable and multivariable MR approaches, while controlling for potential confounders such as income, smoking, alcohol consumption, and various psychiatric diagnoses and symptoms.

Existing literature extensively documents the association between mental health, EA, and CP. High rates of suicidal ideation among college students and the link between prior suicide attempts and reduced academic performance demonstrate these connections. Observational studies suggest that lower EA is a risk factor for suicidal behavior across the life course, contributing to socioeconomic disparities in mortality. However, the relationship is complex and confounded by CP, which is strongly associated with both EA and suicidal behavior. Inferring causality from observational data is challenging due to reverse causation and residual confounding from factors such as socioeconomic status, cultural influences, psychiatric disorders, and behavioral mediators. The study acknowledges the limitations of observational studies in establishing causality and highlights the need for MR analysis to disentangle these intricate relationships.

This study employed a bidirectional and multivariable Mendelian randomization (MR) design. The researchers utilized summary statistics from large genome-wide association studies (GWAS) of EA, CP, and suicide attempt risk, encompassing a combined sample size of more than 815,000 white individuals of European ancestry. Instruments for EA were extracted from a large SSGAC GWAS meta-analysis, while instruments for CP came from a combined SSGAC and UK Biobank GWAS. The suicide attempt data was sourced from hospital-based records. The SNPs selected for each trait met a genome-wide significance threshold (p < 5 × 10⁻⁸). The analysis involved multiple stages. First, single-variable MR (SMR) was used to assess the individual effects of EA and CP on suicide attempt risk. Next, bidirectional SMR was performed to investigate whether suicide attempts influenced EA or CP. Finally, multivariable MR (MVMR) analyses were conducted to isolate the independent effects of EA and CP on suicide attempt risk, controlling for one another and other potential confounding factors. These confounders included household income, tobacco smoking, alcohol consumption, and self-reported symptoms of anxiety and depression, along with several psychiatric diagnoses (major depressive disorder, schizophrenia, and bipolar disorder). Multiple MR methods (IVW, MR Egger, Weighted Median, Weighted Mode, MR PRESSO) were used to test the robustness of the findings across different assumptions regarding the presence of pleiotropy (when a genetic variant influences more than one trait). All analyses were performed using publicly available summary statistics; thus, no new ethical approval was needed.

Single-variable MR analysis indicated that both higher EA and higher CP were associated with a reduced risk of suicide attempts (EA: OR per SD increase = 0.412–0.666, P = 1.70 × 10⁻⁷; CP: OR per SD increase = 0.572–0.885, P = 0.020). Bidirectional MR analyses revealed no evidence that suicide attempts causally influenced EA or CP. The key findings emerged from the multivariable MR (MVMR) analyses, which examined the independent effects of EA and CP while controlling for each other. These analyses demonstrated that EA had a significant independent protective effect against suicide attempts (OR = 0.206–0.568, P = 1.61 × 10⁻⁴), whereas CP showed no independent effect. Further MVMR analyses, incorporating additional behavioral and psychiatric mediators, largely confirmed the independent effect of EA (OR = 0.421–0.696, P = 3.33 × 10⁻¹⁰), with CP still exhibiting no significant association. The consistency of results across different MR methods enhanced confidence in the causal inference. The results strongly suggest that higher EA causally reduces the risk of suicide attempts in white individuals of European ancestry, independent of the effects of CP and after accounting for various confounding factors.

The study's findings provide robust evidence for a causal link between educational attainment and reduced risk of suicide attempts. The independent effect of EA remained significant even after controlling for CP and numerous potential confounders, highlighting its substantial impact on suicide risk. The lack of independent effect for CP suggests that its influence on suicide attempts might be largely mediated through EA. These results have significant implications for public health policy and suicide prevention programs, emphasizing the potential importance of educational interventions and initiatives aimed at increasing educational attainment. The results should inform strategies focused on improving educational opportunities and social support systems to decrease suicide risk.

This large-scale Mendelian randomization study provides strong evidence suggesting a causal relationship between higher educational attainment and a reduced risk of suicide attempts, independent of cognitive performance and several other confounding factors. These results highlight the potential importance of educational interventions in suicide prevention efforts. Future research should explore the generalizability of these findings to diverse populations, including individuals from different ethnic backgrounds and socioeconomic statuses. Further investigation into the underlying mechanisms through which EA reduces suicide risk is also warranted.

The study's limitations include the focus on a specific population (white individuals of European ancestry), limiting the generalizability of the findings. The use of summary-level data from GWAS also restricts the ability to examine gene-environment interactions or other complexities in the underlying causal pathways. Relying on hospital-based records of suicide attempts may also lead to some degree of underreporting or selection bias.