21 Novembre 2026

Psicologia Punto One | Psicologia e Psicologi nel Web

Menu

Menu

SCHEDA SCIENTIFICA2026Altra ricerca

Genetic Liability to Opioid Use Disorder and Dementia-Related Biological Pathways: A Mendelian Randomisation Study.

Sara Javidnia, Héléne Toinét Cronjé, Ville Karhunen, Dipender Gill, Steven Bell, Joseph D Deak, Daniel Levey, Rachel L Kember, Henry R Kranzler, Stephen Burgess, Joel Gelernter, Klaus P Ebmeier, Anya Topiwala

International journal of geriatric psychiatry · 2026

Abstract originale

OBJECTIVES: Opioid use disorder (OUD) is increasingly recognised as a risk factor for cognitive impairment and dementia, but the biological mechanisms underlying this association remain unclear. We aimed to investigate the involvement of established neurodegenerative, vascular, or inflammatory pathways. METHODS: We performed a two-sample Mendelian randomisation study using summary-level genetic data from nine genome-wide association studies of individuals predominantly of European ancestry. Genetic instruments for OUD liability were used to estimate associations with neurodegenerative, vascular, and inflammatory phenotypes, including amyloid-β, tau, cerebral amyloid angiopathy, cerebral microinfarcts, and C-reactive protein (CRP). Causal effects were estimated using inverse-variance weighted Mendelian randomisation, with replication and Bayesian colocalisation analyses used to further evaluate the findings. RESULTS: There was little evidence that genetic liability to OUD was associated with amyloid-β, tau, or vascular phenotypes. A modest positive association was observed with CRP (β = 0.11, 95% CI 0.04-0.18, p = 2 × 10-3). This association was replicated in an independent dataset (β = 0.12, 95% CI 0.05-0.19, p = 1.59 × 10-3). A nominal inverse association was also observed with IL-15 (β = -0.23, 95% CI -0.43 to -0.03, p = 2.2 × 10-2), although colocalisation did not support a shared causal variant at implicated loci. CONCLUSIONS: Genetic liability to OUD showed little evidence of association with the neurodegenerative or vascular phenotypes examined; however, these findings do not exclude the involvement of Alzheimer's disease or cerebrovascular pathways, particularly given the limited power and the use of only three genetic instruments. The observed inflammatory associations should be considered hypothesis-generating, particularly given the lack of colocalisation support, and require further investigation.

Parole chiave

HumansMendelian Randomization AnalysisGenome-Wide Association StudyOpioid-Related DisordersGenetic Predisposition to DiseaseDementiatau ProteinsC-Reactive ProteinAmyloid beta-PeptidesC‐reactive proteindementiainflammationmendelian randomisationneurodegenerationopioid use disorder
Leggi il testo integrale originale

Testo estratto dalla pubblicazione: sono esclusi immagini, media, supplementi e parti soggette a diritti distinti.

Genetic Liability to Opioid Use Disorder and Dementia‐Related Biological Pathways: A Mendelian Randomisation Study

ABSTRACT

Objectives

Opioid use disorder (OUD) is increasingly recognised as a risk factor for cognitive impairment and dementia, but the biological mechanisms underlying this association remain unclear. We aimed to investigate the involvement of established neurodegenerative, vascular, or inflammatory pathways.

Methods

We performed a two‐sample Mendelian randomisation study using summary‐level genetic data from nine genome‐wide association studies of individuals predominantly of European ancestry. Genetic instruments for OUD liability were used to estimate associations with neurodegenerative, vascular, and inflammatory phenotypes, including amyloid‐β, tau, cerebral amyloid angiopathy, cerebral microinfarcts, and C‐reactive protein (CRP). Causal effects were estimated using inverse‐variance weighted Mendelian randomisation, with replication and Bayesian colocalisation analyses used to further evaluate the findings.

Results

There was little evidence that genetic liability to OUD was associated with amyloid‐β, tau, or vascular phenotypes. A modest positive association was observed with CRP (β = 0.11, 95% CI 0.04–0.18, p = 2 × 10−3). This association was replicated in an independent dataset (β = 0.12, 95% CI 0.05–0.19, p = 1.59 × 10−3). A nominal inverse association was also observed with IL‐15 (β = −0.23, 95% CI −0.43 to −0.03, p = 2.2 × 10−2), although colocalisation did not support a shared causal variant at implicated loci.

Conclusions

Genetic liability to OUD showed little evidence of association with the neurodegenerative or vascular phenotypes examined; however, these findings do not exclude the involvement of Alzheimer's disease or cerebrovascular pathways, particularly given the limited power and the use of only three genetic instruments. The observed inflammatory associations should be considered hypothesis‐generating, particularly given the lack of colocalisation support, and require further investigation.

Key Points

Genetic liability to opioid use disorder showed a modest association with higher circulating C‐reactive protein levels.

No evidence was found for associations with major neurodegenerative or vascular biomarkers linked to dementia, although these findings do not exclude Alzheimer's disease or cerebrovascular pathways, particularly given limited power.

The inflammatory findings are hypothesis‐generating and should be interpreted cautiously given the use of only three genetic instruments and the lack of colocalisation support.

1Introduction

Opioid use disorder (OUD) is an escalating public health concern with significant implications for brain health. Individuals with OUD have substantially elevated risk of cognitive impairment and dementia, and recent Mendelian randomisation (MR) analyses support a potential causal relationship with all‐cause dementia [1, 2]. However, the biological mechanisms underlying this association remain unclear, limiting opportunities for targeted prevention and intervention.

Chronic opioid exposure has been associated with structural brain changes, including reductions in hippocampal and white matter volumes, alongside decline in fluid intelligence and executive functioning [3]. Opioids may contribute to neurodegeneration through direct receptor‐mediated effects, and indirectly via co‐occurring factors such as chronic pain, psychological stress, and altered neuroinflammatory signalling [4]. However, it remains unclear whether these changes represent causal pathways to dementia or downstream consequences of OUD and its comorbidities.

There is limited evidence on whether OUD influences established neuropathological processes underlying dementia, including amyloid beta and tau pathology, or vascular mechanisms, such as small vessel disease. The potential role of systemic inflammation, which has been implicated in both OUD and dementia [5], also remains uncertain.

To address this gap, we used two‐sample Mendelian randomisation and Bayesian colocalisation to investigate whether genetic liability to OUD is associated with dementia‐related neurodegenerative, vascular, and inflammatory traits, including amyloid beta, tau, cerebral amyloid angiopathy, microinfarcts, and C‐reactive protein (CRP) [6, 7, 8, 9, 10]. By evaluating these pathways, we sought to determine whether human genetic evidence supports a role for OUD in increasing dementia risk through its potential effects on established dementia‐related mechanisms.

2Materials and Methods

2.1Sample and Data Sources

We used publicly available summary statistics from genome‐wide association studies (GWASs) conducted in predominantly European ancestry populations. Sample sizes varied across the individual exposure and outcome GWASs and are reported in Supporting Information S2: Tables 1 and 2.

2.2Opioid Use Disorder Trait

Genetic variants associated with OUD were obtained from a large GWAS meta‐analysis comprising 7 cohorts (n = 15,251 cases, n = 538,935 controls) [11] (Supporting Information S2: Table 2). Lifetime OUD was identified from International Classification of Diseases (ICD‐9/10) codes in linked electronic health records (EHR) (Supporting Information S2: Table 3). Individuals without an OUD diagnosis were classified as controls [1, 11]. Three independent single‐nucleotide polymorphisms (SNPs) (r
2 < 0.1) reached genome‐wide significance (p < 5 × 10−8) and were used as instrumental variables in the MR analysis: rs79704991 and rs1799971, located in the OPRM1 gene, and rs11372849, located in the FURIN gene (Supporting Information S2: Table 4). Where rs11372849 was not available in the outcome dataset, a proxy SNP in high linkage disequilibrium (rs17514846; r
2 > 0.8) was used instead [1]. Pairwise linkage disequilibrium between the genetic instruments was assessed. The two OPRM1 variants (rs79704991 and rs1799971) showed low LD (r
2 = 0.02), while rs11372849 in FURIN is located on a different chromosome. Given the low correlation between the OPRM1 instruments, all variants were treated as independent in the IVW analyses; residual LD was not incorporated into the IVW variance calculation.

2.3Neurobiological and Inflammatory Biomarkers

Genetic associations for neuroendophenotypes and circulating inflammatory and neurodegeneration‐related biomarkers were obtained from previously published GWAS studies. These included 11 neuropathological traits (n = 7804) [6] (Supporting Information S2: Table 1); plasma tau levels (n = 14,721) [8]; tau and amyloid‐beta levels in cerebrospinal fluid (CSF) (n = 13,116) [7]; plasma neurofilament light chain (NfL) levels measured using bead‐based immunoassays (n = 18,532) and antibody‐based Olink proteomic measurements (n = 49,682) [12, 13]; and Olink‐derived measures of circulating protein abundance in UK Biobank participants (n ≤ 54,219) [13], including Glial Fibrillary Acidic Protein (GFAP), Amyloid Precursor Protein (APP), Brain‐Derived Neurotrophic Factor (BDNF), Microtubule‐Associated Protein Tau (MAPT), interleukin‐4 (IL‐4), and interleukin‐15 (IL‐15) [13]. Additional inflammatory biomarkers included interleukin‐1α (IL‐1α; n = 47,745) [14], interleukin‐6 (IL‐6; n = 74,679), interleukin‐1β (IL‐1β; n = 38,868), interleukin‐18 (IL‐18; n = 70,126), and (Monocyte Chemoattractant Protein‐1, MCP‐1; n = 74,783) [15]; as well as C‐reactive protein (CRP) levels from primary and replication GWAS datasets (n = 575,531 and n = 319,548, respectively) [9, 10]. The scale of the MR estimates was determined by the scale of the SNP–exposure and SNP–outcome associations in the respective GWASs. OUD was analysed as a binary exposure, with SNP–OUD associations expressed on the log‐odds scale. Outcomes comprised continuous, ordinal, and binary phenotypes, with measurement scales, transformations, and standardisation varying across GWASs. Accordingly, MR estimates were interpreted on the scale of the respective outcome per unit increase in the log‐odds of genetic liability to OUD. The phenotype type, measurement or transformation, standardisation, and scale of the SNP–outcome associations for each outcome are detailed in Supporting Information S2: Table 1.

Potential sample overlap between the OUD exposure GWAS and each outcome GWAS was assessed based on the contributing cohorts reported in the source studies. Potential overlap was identified for the CRP replication analysis, as FinnGen contributed to both the OUD GWAS and the CRP dataset [10]. The exact number of overlapping participants could not be quantified from the publicly available summary‐level data. Potential cohort overlap for all outcome GWASs is reported in Supporting Information S2: Table 1.

2.4Genetic Analysis

Two‐sample MR was employed to evaluate the association between genetic liability to OUD and neuroendophenotype or inflammatory cytokine outcomes. The inverse variance weighted (IVW) method was used as the primary analytical approach. Analyses were conducted using the TwoSampleMR (version 0.6.3) and MendelianRandomization packages in R. Additional sensitivity analyses, including Cochran's Q tests for heterogeneity and leave‐one‐out analyses, were conducted using TwoSampleMR version 0.6.22. Mendelian randomisation analyses were conducted under the three‐core instrumental variable assumptions. First, the relevance assumption requires that the genetic instruments are robustly associated with the exposure (OUD). Instruments were selected based on genome‐wide significant associations with OUD, and instrument strength was assessed using the proportion of variance explained (R
2) and F‐statistics, with F > 10 considered indicative of adequate instrument strength. Second, the independence assumption requires that the genetic instruments are independent of factors that confound the association between OUD and the outcomes. Third, the exclusion restriction assumption requires that the genetic instruments influence the outcomes only through their effect on OUD and not through alternative pathways (horizontal pleiotropy). Pairwise linkage disequilibrium between instruments was assessed to confirm their independence, and heterogeneity across SNP‐specific causal estimates was evaluated using Cochran's Q statistic. Leave‐one‐out sensitivity analyses were additionally performed by sequentially excluding each genetic instrument to assess whether the overall MR estimates were disproportionately influenced by any single variant. Leave‐one‐out analysis was not performed for IL‐1β because only two genetic instruments were available. Given the limited number of genetic instruments, the sensitivity analyses were interpreted cautiously. Heterogeneity and leave‐one‐out results are presented in Supporting Information S2: Tables 9 and 10, respectively. Replication and colocalisation analyses were undertaken to provide additional evidence supporting the robustness of the findings [16]. Prior to analysis, SNP‐exposure and SNP‐outcome associations were harmonised to ensure alignment on the effect allele. To assess potential confounding due to linkage disequilibrium, Bayesian colocalisation analysis was performed using the coloc package (version 5.2.3) [16]. Multiple testing correction was performed using false discovery rate (FDR) (Benjamini–Hochberg, 5%). The inflammatory biomarkers and neuroendophenotypes were treated as separate families of hypotheses because they represented distinct outcome domains addressing separate research questions. Benjamini–Hochberg correction was therefore applied separately across the eight inflammatory biomarkers and the 20 neuroendophenotypes, with the false discovery rate controlled at 5% within each family. Replication analyses were additionally performed for CRP using an independent GWAS dataset [10].

We used Bayesian colocalisation analysis to estimate posterior probabilities for five mutually exclusive hypotheses: no association with either trait (PP.H0), association with OUD liability only (PP.H1), association with the outcome only (PP.H2), association with both traits through distinct causal variants (PP.H3), and association with both traits through a shared causal variant (PP.H4) [17]. We first assessed evidence that both traits were associated within the locus using PP.H3 + PP.H4 > 0.5. Among loci meeting this criterion, support for a shared rather than distinct causal variant was summarised using PP.H4/(PP.H4 + PP.H3). Values > 0.5 were interpreted as greater posterior support for colocalisation than for distinct causal variants, whereas values < 0.5 were interpreted as greater support for distinct association signals. Colocalisation analyses were performed under the standard coloc assumption of a single causal variant per trait within each genomic region, and results were interpreted accordingly. This study was not preregistered.

3Results

Three independent variants in OPRM1 and FURIN associated with OUD at genome‐wide significance were selected as instrumental variables for MR analyses. Using the inverse‐variance weighted approach, greater genetic liability to OUD was associated with higher circulating CRP levels (inverse‐variance weighted [IVW] β = 0.11, 95% CI [0.04, 0.18], p = 2.21 × 10−3) (Figure 1) (Supporting Information S2: Table 5). Replication analyses using an independent CRP dataset demonstrated a consistent direction of effect and slightly stronger association estimates (IVW β = 0.12, 95% CI [0.05, 0.19], p = 1.59 × 10−3) (Supporting Information S2: Table 7). There was nominal evidence of an inverse association between genetic liability to OUD and circulating IL‐15 levels (IVW β = −0.23, 95% CI −0.43 to −0.03; p = 0.022), although this did not survive correction for multiple testing at the prespecified 5% FDR threshold (Supporting Information S2: Table 5).

For further details on the inflammatory biomarker analyses, see Supporting Information S1: Figure 3.

To assess whether this association reflected a shared underlying causal variant, we performed Bayesian colocalisation analyses at the genome‐wide significant loci. Colocalisation analyses did not support a shared association between OUD and CRP at either locus. For the primary CRP analysis, PP.H3 + PP.H4 was 0.196 at OPRM1 and 0.040 at FURIN, below the predefined threshold of > 0.5. Corresponding values were lower in the replication analysis. Across these analyses, PP.H1, representing association with OUD alone, had the highest posterior probability (Figure 2). This suggests that the MR finding is unlikely to be explained by a single shared causal variant at these loci. Similar findings were observed for IL‐15 and the independent CRP replication dataset, with limited evidence supporting colocalisation across analyses Supporting Information S2: Table 8 and Supporting Information S1: Figures 1 and 2.

Aside from the CRP association described above, no associations across the broader panel of inflammatory markers and neuroendophenotypes reached statistical significance after correction for multiple testing (Supporting Information S2: Tables 5 and 6; Figure 3). Whilst weak inverse associations between OUD liability and plasma neurofilament light chain levels (IVW β = −0.28, 95% CI [−0.49, −0.07], p = 9.99 × 10−3), and cerebral amyloid angiopathy (CAA) burden (IVW β = −1.34, 95% CI [−2.39, −0.29], p = 0.013) were observed, these did not survive false discovery rate (FDR) correction and therefore should be interpreted with caution (Supporting Information S2: Table 6; Figure 3).

Further details of the neuroendophenotype analyses are provided in Supporting Information S1: Figure 4.

4Discussion

We used Mendelian randomisation to investigate whether genetic liability to OUD influences inflammatory and neurodegenerative pathways. We found evidence that greater genetically predicted liability to OUD is associated with higher circulating levels of C‐reactive protein (CRP), supporting a link with systemic inflammation. This association was replicated in an independent dataset [10], strengthening the robustness of the finding. We also observed a nominal inverse association between genetic liability to OUD and circulating interleukin‐15 (IL‐15) levels, however, this finding did not survive correction for multiple testing and should be interpreted cautiously.

The absence of colocalisation and the limited number of OUD genetic instruments preclude conclusions regarding shared causal variants or a causal mediating role for CRP. These findings should therefore be considered hypothesis‐generating and require confirmation in larger genetic studies. Beyond CRP, we found little evidence for causal effects of instrumented OUD liability on other outcomes, including Alzheimer's disease–related biomarkers (tau, amyloid‐β), markers of neuroaxonal damage or astroglia activation (NfL, GFAP), or vascular pathology. Nominal associations with NfL and cerebral amyloid angiopathy did not survive correction for multiple testing. These findings should be interpreted in the context of previous literature. While observational studies have reported associations between OUD and dementia risk, the underlying biological mechanisms remain unclear [1, 2]. Our recent MR study provided evidence supporting a causal relationship between OUD and all‐cause dementia [1], but did not find associations with stroke, despite hypotheses implicating cerebrovascular pathways. Together, this suggests mechanisms other than overt vascular disease may be important.

The association with CRP is consistent with clinical studies reporting elevated inflammatory markers in individuals with OUD, which have been linked to cognitive impairment [18]. CRP may be particularly sensitive in this context, reflecting its role as a stable, downstream marker of systemic inflammation, in contrast to more transient cytokines and their relatively underpowered genetic instruments. As a non‐specific marker integrating multiple inflammatory pathways, CRP may capture broader immune dysregulation associated with OUD. CRP GWAS are also substantially larger than those for individual cytokines, providing greater statistical power. This may also suggest that genetic liability to OUD primarily influences broader inflammatory processes rather than specific cytokine pathways. Although IL‐15 also showed a nominally significant association, this finding did not survive multiple testing correction and requires further testing in larger datasets. Several biological mechanisms could underlie this relationship. Chronic opioid exposure is known to activate peripheral and central immune responses, increase pro‐inflammatory signalling, and regularly disrupt systems that include the hypothalamic‐pituitary‐adrenal axis. Opioids may also impair gut barrier function, promoting systemic immune activation. Systemic inflammation, in turn, may influence brain health through effects on the blood–brain barrier and neuroinflammatory pathways, providing a plausible link between OUD, cognitive impairment, and dementia risk [5].

In contrast, we found little evidence linking OUD liability to specific neurodegenerative biomarkers. Previous studies in this area have been inconsistent and often limited by small samples or heterogeneous populations [19]. Our null findings may reflect limited statistical power in available GWAS datasets, particularly for neuropathological markers, and do not exclude more subtle or indirect effects. Larger and more specific datasets, particularly in populations with OUD, will be needed to clarify these relationships.

Key strengths of our study include the use of a causal inference framework such as Mendelian randomisation, which is less susceptible to bias from environmental confounding and reverse causality. We also integrated multiple large‐scale GWAS datasets across biologically relevant inflammatory and neurodegenerative domains. However, several limitations should be noted. Statistical power was limited both for OUD and for some outcomes, particularly neurodegenerative markers. It is also important to note that our exposure represents genetic liability to OUD rather than the clinical phenotype itself. Genetic liability and the clinical disorder may reflect distinct biological mechanisms, and our findings should therefore be interpreted in the context of genetic predisposition rather than manifest disorder. Additionally, several biomarker datasets were derived from indirect plasma protein measurements obtained using different assay platforms, including Olink proteomic and bead‐based immunoassays, which may vary in sensitivity and specificity across proteins and could contribute to measurement heterogeneity. Potential participant overlap between the OUD exposure GWAS and the FinnGen CRP replication dataset could bias two‐sample MR estimates, particularly in the presence of weak instruments; however, all OUD instruments had F‐statistics > 30, reducing the potential for substantial weak‐instrument bias. Nevertheless, this potential overlap should be considered when interpreting the CRP replication findings. Null colocalisation findings across CRP and IL‐15 analyses should also be interpreted cautiously given the assumptions underlying standard coloc models. Residual pleiotropy cannot be excluded, and variation in phenotype definitions across contributing studies may introduce measurement error. Given limited sample sizes in other populations, analyses were restricted to predominantly European ancestry populations, limiting generalizability. Finally, the lack of biomarker data specifically from individuals with OUD may reduce the specificity of our findings.

5Conclusion

Genetic liability to OUD was associated with elevated circulating CRP levels, with this association replicated in an independent dataset, raising the hypothesis that systemic inflammation may be relevant in the relationship between OUD and dementia risk. A nominal association with IL‐15 provides an additional signal that warrants replication. However, we found no evidence for associations with the neurodegenerative or vascular phenotypes examined, although these findings do not exclude Alzheimer's disease or cerebrovascular pathways, particularly given limited power and the use of only three genetic instruments. The inflammatory findings should therefore be considered hypothesis‐generating, particularly given the lack of colocalisation support. Further large‐scale studies, particularly those incorporating well‐characterised OUD populations and longitudinal biomarker data, are needed to clarify the underlying mechanisms of systemic inflammation associated with OUD and inform prevention strategies.

Ethics Statement

This study used publicly available summary‐level GWAS data from studies with existing ethical approval and participant informed consent.

Conflicts of Interest

D.G. is the Chief Executive Officer of Sequoia Genetics, a private limited company that works with investors, pharma, biotech, and academia by performing research that leverages genetic data to help inform drug discovery and development. D.G. has financial interests in several biotechnology companies. H.T.C is employed by Sequoia Genetics. SBu is an employee of Sequoia Genetics, a private limited company that works with investors, pharma, biotech, and academia by performing research that leverages genetic data to help inform drug discovery and development. Dr. Kranzler is a member of advisory boards for Altimmune and Clearmind Medicine; a consultant to Sobrera Pharmaceuticals, Altimmune, Lilly, Ribocure, and Boehringer‐Ingelheim; and the recipient of research funding and medication supplies for an investigator‐initiated study from Alkermes and company‐initiated studies by Altimmune and Lilly.

Supporting information

References

1S. Javidnia , J. M. Roe , V. Karhunen , et al., “Opioid Use Disorder and Dementia Risk: Evidence From Observational and Genetic Analyses in Diverse Ancestry Cohorts ,” Alzheimer's and Dementia
22 , no. 5 (2026 ): e71418, 10.1002/alz.71418 .PMC13240559 42170857

2F. Qeadan , A. McCunn , B. Tingey , et al., “Exploring the Association Between Opioid Use Disorder and Alzheimer’s Disease and Dementia Among a National Sample of the U.S. Population ,” Journal of Alzheimer's Disease
96 , no. 1 (2023 ): 229 –244 , 10.3233/JAD-230714 .37742654

3M. T. Bennett , Y. Modna , and D. K. Shah , “The Impact of Opioid Drugs on Memory and Other Cognitive Functions: A Review ,” Journal of Biosciences and Medicines
12 , no. 4 (2024 ): 264 –287 , 10.4236/jbm.2024.124021 .

4K. C. Reeves , N. Shah , B. Muñoz , and B. K. Atwood , “Opioid Receptor‐Mediated Regulation of Neurotransmission in the Brain ,” Frontiers in Molecular Neuroscience
15 (2022 ): 919773, 10.3389/fnmol.2022.919773 .PMC9242007 35782382

5E. R. Butelman , R. Z. Goldstein , C. A. Nwaneshiudu , et al., “Neuroimmune Mechanisms of Opioid Use Disorder and Recovery: Translatability to Human Studies, and Future Research Directions ,” Neuroscience
528 (2023 ): 102 –116 , 10.1016/j.neuroscience.2023.07.031 .PMC10720374 37562536

6L. M. P. Shade , Y. Katsumata , E. L. Abner , et al., “GWAS of Multiple Neuropathology Endophenotypes Identifies New Risk Loci and Provides Insights Into the Genetic Risk of Dementia ,” Nature Genetics
56 , no. 11 (2024 ): 2407 –2421 , 10.1038/s41588-024-01939-9 .PMC11549054 39379761

7I. E. Jansen , S. J. van der Lee , D. Gomez‐Fonseca , et al., “Genome‐Wide Meta‐Analysis for Alzheimer’s Disease Cerebrospinal Fluid Biomarkers ,” Acta Neuropathologica
144 , no. 5 (2022 ): 821 –842 , 10.1007/s00401-022-02454-z .PMC9547780 36066633

8C. Sarnowski , M. Ghanbari , J. C. Bis , et al., “Meta‐Analysis of Genome‐Wide Association Studies Identifies Ancestry‐Specific Associations Underlying Circulating Total Tau Levels ,” Communications Biology
5 , no. 1 (2022 ): 336 , 10.1038/s42003-022-03287-y .PMC8993877 35396452

9S. Said , R. Pazoki , V. Karhunen , et al., “Genetic Analysis of Over Half a Million People Characterises C‐Reactive Protein Loci ,” Nature Communications
13 , no. 1 (2022 ): 2198, 10.1038/s41467-022-29650-5 .PMC9033829 35459240

10M. I. Kurki , J. Karjalainen , P. Palta , et al., “FinnGen Provides Genetic Insights From a Well‐Phenotyped Isolated Population ,” Nature
613 , no. 7944 (2023 ): 508 –518 , 10.1038/s41586-022-05473-8 .PMC9849126 36653562

11J. D. Deak , H. Zhou , M. Galimberti , et al., “Genome‐Wide Association Study in Individuals of European and African Ancestry and Multi‐Trait Analysis of Opioid Use Disorder Identifies 19 Independent Genome‐Wide Significant Risk Loci ,” Molecular Psychiatry
27 , no. 10 (2022 ): 3970 –3979 , 10.1038/s41380-022-01709-1 .PMC9718667 35879402

12S. Ahmad , M. A. Imtiaz , A. Mishra , et al., “Genome‐Wide Association Study Meta‐Analysis of Neurofilament Light (NfL) Levels in Blood Reveals Novel Loci Related to Neurodegeneration ,” Communications Biology
7 , no. 1 (2024 ): 1103, 10.1038/s42003-024-06804-3 .PMC11385583 39251807

13B. B. Sun , J. Chiou , M. Traylor , et al., “Plasma Proteomic Associations With Genetics and Health in the UK Biobank ,” Nature
622 , no. 7982 (2023 ): 329 –338 , 10.1038/s41586-023-06592-6 .PMC10567551 37794186

14H. Loya , G. Kalantzis , F. Cooper , and P. F. Palamara , “A Scalable Variational Inference Approach for Increased Mixed‐Model Association Power ,” Nature Genetics
57 , no. 2 (2025 ): 461 –468 , 10.1038/s41588-024-02044-7 .PMC11821521 39789286

15M. J. Konieczny , M. Omarov , L. Zhang , et al., “The Genomic Architecture of Circulating Cytokine Levels Points to Drug Targets for Immune‐Related Diseases ,” Communications Biology
8 , no. 1 (2025 ): 34 , 10.1038/s42003-025-07453-w .PMC11724035 39794498

16C. Wallace , “Eliciting Priors and Relaxing the Single Causal Variant Assumption in Colocalisation Analyses ,” PLoS Genetics
16 , no. 4 (2020 ): e1008720, 10.1371/journal.pgen.1008720 .PMC7192519 32310995

17V. Zuber , N. F. Grinberg , D. Gill , et al., “Combining Evidence From Mendelian Randomization and Colocalization: Review and Comparison of Approaches ,” American Journal of Human Genetics
109 , no. 5 (2022 ): 767 –782 , 10.1016/j.ajhg.2022.04.001 .PMC7612737 35452592

18T. Y. Wang , R. B. Lu , S. Y. Lee , et al., “Association Between Inflammatory Cytokines, Executive Function, and Substance Use in Patients With Opioid Use Disorder and Amphetamine‐Type Stimulants Use Disorder ,” International Journal of Neuropsychopharmacology
26 , no. 1 (2023 ): 42 –51 , 10.1093/ijnp/pyac069 .PMC9850661 36181736

19K. Agarwal , P. Manza , M. Chapman , et al., “Inflammatory Markers in Substance Use and Mood Disorders: A Neuroimaging Perspective ,” Frontiers in Psychiatry
13 (2022 ): 863734, 10.3389/fpsyt.2022.863734 .PMC9086785 35558424

Come citare questo studio

Sara Javidnia, Héléne Toinét Cronjé, Ville Karhunen, Dipender Gill, Steven Bell, Joseph D Deak, Daniel Levey, Rachel L Kember, Henry R Kranzler, Stephen Burgess, Joel Gelernter, Klaus P Ebmeier, Anya Topiwala. (2026). Genetic Liability to Opioid Use Disorder and Dementia-Related Biological Pathways: A Mendelian Randomisation Study.. International journal of geriatric psychiatry. DOI: 10.1002/gps.70270

Consulta la pubblicazione originale ↗