Causal effects of lung function on brain cortical structure and subcortical structure: a two-sample univariate and multivariate Mendelian randomization study

doi:10.21203/rs.3.rs-4963908/v1

Download PDF

Article

Causal effects of lung function on brain cortical structure and subcortical structure: a two-sample univariate and multivariate Mendelian randomization study

https://doi.org/10.21203/rs.3.rs-4963908/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Lung function has been increasingly linked to overall health, including brain health, prompting the investigation into the causal relationships between lung function and brain structures. This study employs Mendelian Randomization (MR) to explore these causal relationships, leveraging genetic variants as proxies to predict the effects of lung function on brain cortical and subcortical structures.

Methods

We conducted univariate and multivariable MR analyses using GWAS summary statistics for lung function (FEV1, FVC, FEV1/FVC) and brain structures from the UK Biobank and ENIGMA consortium. Our analyses included five MR methods—IVW, MR-Egger, weighted median, weighted mode, and simple mode—to ensure robust causal inference. Multivariable MR (MVMR) analyses were performed to adjust for potential confounders like smoking and education. Sensitivity analyses were performed to confirm the stability of our results, and we applied FDR correction for multiple comparisons.

Results

The univariate MR analysis revealed significant associations between lung function and brain structures. Higher FEV1 was associated with increased global cortical volume (β = 4428.037, SE = 610.453, p < 0.0001) and supramarginal thickness (β = 43.613, SE = 13.218, p = 0.001). FVC was similarly associated with increased global cortical volume (β = 3650.674, SE = 576.736, p < 0.0001) and parsopercularis thickness (β = 0.013, SE = 0.003, p < 0.0001). Multivariable MR confirmed these associations, even after adjusting for smoking and education. Significant associations persisted in subcortical regions, with higher FEV1 and FVC linked to increased brainstem volume (FEV1: β = 0.226, SE = 0.049, p < 0.0001; FVC: β = 0.203, SE = 0.044, p < 0.0001) and amygdala volume for FEV1/FVC (β = 0.075, SE = 0.025, p = 0.003). Sensitivity and pleiotropy analyses indicated no significant heterogeneity or horizontal pleiotropy, confirming the robustness of the results.

Conclusion

Our study provides robust evidence of a causal relationship between lung function and brain structure, emphasizing the protective effects of better respiratory health on brain integrity. However, the reliance on European GWAS data limits generalizability, and some associations did not survive stringent correction. Future research should incorporate diverse populations and explore underlying mechanisms to validate and extend these findings.

Biological sciences/Neuroscience

Health sciences/Diseases

Health sciences/Neurology

Lung function

Cortical structure

Subcortical structure

Causal relationship

Mendelian randomization

Genome-wide association studies

Entering the 21st century, rapid economic development has brought about significant industrial and environmental pollution, imposing a substantial burden on human lung health¹. Epidemiological data indicate that in 2019, there were approximately 212.3 million individuals suffering from chronic obstructive pulmonary disease (COPD) worldwide², leading to 3.23 million deaths and making it the third leading cause of death globally³. COPD severely impacts patients' quality of life, causing chronic respiratory symptoms such as persistent cough, excessive mucus production, and breathlessness⁴. It also increases the risk of acute respiratory infections and cardiovascular diseases, contributing to higher morbidity and mortality rates^5,6. Unfortunately, once patients progress to the COPD stage, the pathological damage to the lungs and respiratory function becomes irreversible⁷. Consequently, there is growing emphasis among researchers on the importance of lung function assessments, including forced expiratory volume in one second (FEV1), forced vital capacity (FVC), and the ratio of FEV1 to FVC (FEV1)⁸. Studies have increasingly demonstrated that these indicators are closely linked to overall human health, particularly brain health, which has led to the emergence of the lung-brain axis concept^9–11.

When reviewing previous studies, numerous investigations have suggested a correlation between COPD and changes in brain structure, including alterations in both gray and white matter in specific brain regions^12,13. Specifically, patients with COPD exhibit abnormal activation in the prefrontal network and perfusion damage in white matter¹⁴. Another study has shown reduced white matter integrity throughout the brain in stable non-hypoxemic COPD patients¹⁵. Moreover, research has demonstrated a 1.1% reduction in normalized gray matter volume in COPD patients, a rate of decline surpassing the typical annual gray matter volume loss observed with aging¹⁶. However, we must acknowledge several limitations of these studies, including their cross-sectional observational nature, small sample sizes, and contradictory findings. These limitations undermine the strength of the current evidence, highlighting the need for more robust and confirmatory studies. Furthermore, research specifically examining the relationship between lung function and brain structural and functional changes is scarce. This gap is clinically significant, as early detection of concomitant brain damage in patients with impaired lung function is crucial for timely intervention and treatment. In recent years, Mendelian Randomization (MR) has emerged as a novel research approach that uses genetic variants as proxies for exposures to predict their effects on outcomes¹⁷. Because genetic variants are randomly assigned at conception, MR can mitigate the effects of confounding factors and reverse causation¹⁸. Additionally, the rise of genome-wide association studies (GWAS) provides an opportunity to use MR to infer the causal effects of lung function on brain structure^19–21.

Therefore, this study aims to explore the impact of lung function on brain structure using publicly available GWAS data and Mendelian Randomization analyses. Our research seeks to identify the effects of lung function on specific brain regions and further substantiate the existence of the lung-brain axis. We hypothesize that a decline in lung function has significant implications for brain health. Early detection of brain damage in patients with impaired lung function is crucial for improving quality of life, reducing disease burden, and promoting overall human health.

Firstly, Table 1 presents the sources of all GWAS data used in our MR analysis. Additionally, Fig. 1 outlines the entire data analysis workflow and the statistical methods employed. All our analyses were conducted in accordance with the STROBE-MR Checklist, as detailed in Supplementary Table S1. Supplementary Table S2 lists the IVs information extracted for FEV1, FVC, FEV1/FVC, smoking, and education. Next, we merged the IVs of the exposure variables with the outcome data. The data included in the final MR analysis are provided in Supplementary Table S3. The results of all MR analyses are displayed in Figs. 2 and 3, and Supplementary Tables S4-S6. Figures 4 and Supplementary Table S7 present the results of the MVMR analysis. Lastly, Supplementary Tables S8-S9 show the results of sensitivity and pleiotropy analyses. In summary, our findings indicate a causal relationship between lung function and brain structure. This causal relationship remains statistically significant even after adjusting for smoking and education.

2.1 Univariate MR results: effects of lung function on cortical structures

The univariate MR analysis demonstrated several associations between lung function metrics and cortical structures (Figs. 2, 3, and Supplementary Tables S4-S5). Before FDR correction, the results suggested potential causal relationships. Specifically, higher FEV1 was associated with increased global cortical volume (β = 4428.037, SE = 610.453, p = 0.000) and supramarginal thickness in the parietal region (β = 43.613, SE = 13.218, p = 0.001). Additionally, higher FEV1 showed a negative association with lingual thickness in the occipital region (β = -26.014, SE = 12.620, p = 0.039) and postcentral thickness in the parietal region (β = -24.787, SE = 12.056, p = 0.040). For FVC, significant positive associations were observed with global cortical volume (β = 3650.674, SE = 576.736, p = 0.000) and several other cortical regions, including the parsopercularis in the frontal lobe (β = 13.329, SE = 6.355, p = 0.036), insula in the temporal lobe (β = 19.261, SE = 6.411, p = 0.003), and inferior parietal region (β = 31.421, SE = 15.370, p = 0.041). Negative associations were found with the cuneus (β = -14.211, SE = 5.444, p = 0.009) and lateral occipital regions (β = -40.397, SE = 14.165, p = 0.004). The FEV1/FVC ratio also showed a significant positive association with the paracentral region in the frontal lobe (β = 13.076, SE = 3.538, p = 0.000). After FDR correction, the significant associations that persisted included FEV1 with global cortical volume (β = 4428.037, SE = 610.453, p = 0.000) and supramarginal thickness in the parietal region (β = 43.613, SE = 13.218, p = 0.001), FVC with global cortical volume (β = 3650.674, SE = 576.736, p = 0.000), and FEV1/FVC ratio with the paracentral region in the frontal lobe (β = 13.076, SE = 3.538, p = 0.000). Further analysis of cortical thickness revealed additional associations. Before FDR correction, FEV1 was significantly associated with increased thickness in the caudal middle frontal region (β = 0.010, SE = 0.004, p = 0.013). FVC was positively associated with thickness in the caudal middle frontal (β = 0.008, SE = 0.004, p = 0.028), parsopercularis (β = 0.013, SE = 0.003, p = 0.0001), and temporal pole regions (β = 0.013, SE = 0.006, p = 0.032), while showing a negative association with entorhinal (β = -0.027, SE = 0.012, p = 0.017) and inferior temporal regions (β = -0.010, SE = 0.004, p = 0.021). After FDR correction, the significant association that persisted was FVC with parsopercularis thickness (β = 0.013, SE = 0.003, p = 0.0001).

2.2 Univariate MR results: effects of lung function on subcortical structures

The univariate MR analysis also revealed significant associations between lung function metrics and subcortical structures (Figs. 2, 3, and Supplementary Tables S6). Before FDR correction, higher FEV1 was associated with increased intracranial volume (β = 0.006, SE = 0.002, p = 0.007) and brainstem volume (β = 0.226, SE = 0.049, p = 0.000004). For FVC, significant positive associations were observed with intracranial volume (β = 0.006, SE = 0.002, p = 0.002) and brainstem volume (β = 0.203, SE = 0.044, p = 0.000004). The FEV1/FVC ratio showed a significant positive association with the amygdala volume (β = 0.075, SE = 0.025, p = 0.003). After FDR correction, the significant associations that persisted included FEV1 with intracranial volume (β = 0.006, SE = 0.002, p = 0.007) and brainstem volume (β = 0.226, SE = 0.049, p = 0.000004), FVC with intracranial volume (β = 0.006, SE = 0.002, p = 0.002) and brainstem volume (β = 0.203, SE = 0.044, p = 0.000004), and FEV1/FVC ratio with amygdala volume (β = 0.075, SE = 0.025, p = 0.003).

2.3 Multivariable MR and sensitivity analysis results

The MVMR analysis, which adjusted for smoking and educational attainment, confirmed significant causal relationships between lung function and various brain structures (Figs. 4, Supplementary Table S7). For cortical structures, higher FEV1 was associated with increased global cortical volume (β = 5032.62, 95% CI = 3546.23 to 6519.02, p = 3.22E-11) and supramarginal thickness (β = 46.83, 95% CI = 19.95 to 73.72, p = 6.40E-04). Higher FVC was similarly associated with global cortical volume (β = 4331.01, 95% CI = 3006.89 to 5655.12, p = 1.45E-10) and insular surface area (β = 21.41, 95% CI = 8.10 to 34.71, p = 0.002). In terms of subcortical structures, higher FEV1 and FVC showed strong associations with brainstem volume (FEV1: β = 0.22, 95% CI = 0.12 to 0.32, p = 1.06E-05; FVC: β = 0.20, 95% CI = 0.12 to 0.29, p = 6.47E-06). The FEV1/FVC ratio was significantly associated with amygdala volume (β = 0.08, 95% CI = 0.04 to 0.13, p = 8.77E-04).

The sensitivity analyses, including Cochran's Q test, MR-PRESSO global test, and MR-Egger intercept, indicated no significant heterogeneity or horizontal pleiotropy across most comparisons (Supplementary Tables S8-S9). For example, Cochran's Q test for SA_total with FEV1 showed no heterogeneity (Q' test p = 0.666; IVW Q' test p = 0.652), and the MR-Egger intercept was not significant (p = 0.188), confirming the robustness of the MR findings. Similar results were observed for other outcomes such as brainstem volume and amygdala volume, reinforcing the validity of our causal inferences between lung function and brain structures.

MR analysis is currently one of the most widely used methods for simulating randomized controlled trials to investigate causal relationships. To our knowledge, this is the first study to explore the causal relationship between lung function and brain cortical and subcortical structures. Our findings indicate a causal link between lung function and brain health. Specifically, better lung function, as measured by FEV1, FVC, and the FEV1/FVC ratio, promotes brain health, including increased brain volume and cortical thickness. Given the increasing environmental pollution that burdens lung function, timely lung function assessments and brain-related functional screening are crucial for promoting overall health and should be prioritized by clinicians. Furthermore, our study reaffirms the existence of the lung-brain axis, providing additional clues for screening potential brain health issues or initiating timely interventions. The implications of our findings underscore the importance of maintaining good respiratory health to support neurological integrity, emphasizing the need for integrated clinical approaches to address both pulmonary and brain health.

Lung function is a critical determinant of overall health and has been linked to a variety of health outcomes²². For instance, impaired lung function has been associated with an increased risk of cardiovascular diseases. Studies have shown that lower FEV1 and FVC are predictive of higher incidence of heart attacks and strokes^23,24. Furthermore, COPD, characterized by persistent airflow limitation, significantly increases the risk of acute respiratory infections and hospitalizations. During the COVID-19 pandemic, the importance of lung function became even more apparent. Patients with compromised lung function, such as those with pre-existing respiratory conditions, experienced worse outcomes when infected with SARS-CoV-2²⁵. Reduced lung function was associated with increased severity of COVID-19, higher rates of intensive care unit admission, and increased mortality²⁶. Additionally, long-term sequelae, known as long COVID, have been observed to impact lung function, leading to prolonged respiratory symptoms and reduced exercise capacity. Impaired lung function also has been linked to adverse cognitive outcomes²⁷. Research indicates that lower FEV1 and FVC are associated with an increased risk of cognitive decline and dementia.²⁸ This highlights the interconnectedness of respiratory health and brain function, underscoring the importance of maintaining good lung function not only for respiratory health but also for cognitive well-being.

Our findings, which indicate a causal relationship between lung function and brain structure, align with and extend existing research on the connection between respiratory health and cognitive function. Current evidence largely consists of cross-sectional studies that have demonstrated associations between impaired lung function and cognitive decline²⁹. For instance, numerous studies have shown that lower FEV1 and FVC are linked to poorer cognitive performance, suggesting that respiratory health may play a crucial role in maintaining cognitive function¹⁴. One notable longitudinal study, the Rush Memory and Aging Project, reported that reduced lung function was associated with a more rapid decline in overall cognitive abilities as well as specific cognitive domains, including episodic memory, semantic memory, working memory, visuospatial ability, and perceptual speed³⁰. This study highlights the potential long-term impact of poor lung function on cognitive health, emphasizing the need for proactive management of respiratory conditions to prevent or mitigate cognitive decline. Additionally, our results are consistent with research indicating that poor lung function is associated with adverse findings on brain imaging³¹. Studies have demonstrated that individuals with reduced lung function often exhibit structural brain changes, such as reduced gray matter volume and white matter integrity³². These imaging findings corroborate the observed cognitive deficits, providing a neuroanatomical basis for the relationship between lung function and cognitive impairment.

The mechanisms through which decreased lung function may lead to brain structure damage are multifaceted and complex. One prominent hypothesis is chronic hypoxia. Impaired lung function can result in reduced oxygen levels in the blood, leading to chronic hypoxia, which in turn can cause mitochondrial dysfunction³³. Mitochondria, the powerhouses of cells, are critical for energy production. Dysfunctional mitochondria can lead to energy deficits in neurons, disrupting neural connectivity and impairing brain function³⁴. Another key mechanism is oxidative stress. Reduced lung function is often accompanied by increased production of reactive oxygen species (ROS), leading to oxidative stress³⁵. This condition damages cellular components, including lipids, proteins, and DNA. Oxidative stress can impair cellular endocytosis, a process crucial for nutrient uptake and waste removal in neurons, ultimately leading to neuronal damage and death³⁶. Lastly, impaired lung function can negatively affect neural connectivity and cerebral blood flow³⁷. Hypoxia and oxidative stress can disrupt the blood-brain barrier, a critical structure that regulates the exchange of substances between the blood and the brain^38,39. This disruption can lead to an influx of harmful substances into the brain, further damaging neural tissues. Additionally, poor lung function can lead to reduced cerebral perfusion, compromising the delivery of oxygen and nutrients to brain tissues⁴⁰. This can impair neurovascular coupling, the process by which blood flow is regulated in response to neural activity, ultimately affecting cognitive functions and brain health^41,42.

Summarizing the results of each study must be done with careful consideration of potential limitations to appropriately apply the findings to clinical practice. Firstly, the GWAS data utilized primarily originates from European populations, which limits the generalizability of our results. Future studies need to incorporate data from diverse ethnic groups to validate our findings. Secondly, some of our results did not survive FDR correction, and we did not employ more stringent Bonferroni correction, which could reduce the reliability of our conclusions. Thirdly, although we conducted multivariable analyses, we could not fully assess the extent and impact of covariates on the observed causal relationships. Lastly, the timing of participants' lung function assessments is unknown, potentially affecting the credibility of our results. In summary, while we have identified a potential causal relationship between lung function and brain structure, larger cohort studies are required to confirm these findings and explore deeper mechanisms to better understand the underlying connections.

Our study provides robust evidence for a causal relationship between lung function and brain structure, highlighting that better respiratory health, as indicated by higher FEV1 and FVC, is associated with increased cortical and subcortical volumes. These findings suggest that maintaining good lung function may have protective effects on brain integrity, potentially informing clinical strategies for early detection and intervention in patients with impaired lung function to prevent or mitigate neurodegenerative changes. However, the study's primary limitation is the potential residual confounding despite the use of MVMR. Future research should aim to explore longitudinal data and incorporate more diverse populations to validate these findings and understand the underlying biological mechanisms linking lung function to brain health.

5.1 Study Design

The study overview and MR analysis workflow are illustrated in Fig. 1. MR must satisfy three key assumptions: (1) Relevance: the IV must be strongly associated with the exposure variable; (2) Independence: the IV must not be associated with any confounders of the exposure-outcome relationship; and (3) Exclusion Restriction: the IV should influence the outcome only through the exposure. For lung function metrics, we selected FEV1, FVC, and the FEV1/FVC ratio as the exposure variables. Brain cortical and subcortical structures were chosen as the outcome variables, with the former including cortical thickness and volume, and the latter focusing on subcortical volume. Considering the associations between lung function, smoking, and education, we also performed multivariable MR (MVMR) analyses to selectively control for education or smoking.

In our MR study, all GWAS data for lung function, smoking, education, and brain structure were derived from publicly available sources and had received ethical approval. Detailed information can be found in Table 1. Therefore, additional ethical approval from our committee was not required for this study.

5.2 Data Sources

5.2.1 Data Sources for FEV1, FVC, and the FEV1/FVC

The GWAS summary statistics for lung function metrics FEV1 and FVC were obtained from the UK Biobank, which includes a sample size of 421,986 individuals⁴³. The UK Biobank is one of the largest repositories of genetic and environmental data, established in the UK between 2006 and 2010. The cohort comprises participants of European ancestry aged 39–72, with 55.6% being female. Lung function assessments were conducted using a Vitalograph Pneumotrac 6800 spirometer (Maids Moreton, UK) under the supervision of trained medical personnel. The data were adjusted for age, age squared, height, and smoking status to ensure accuracy. For the FEV1/FVC ratio, the GWAS data were extracted from another study with a sample size of 321,047 individuals⁴⁴. This study was also conducted using data from the UK Biobank. Detailed information about the GWAS data sources is provided in Table 1.

5.2.2 Data Sources for cerebral cortex SA, TH, and subcortical structure

The GWAS summary statistics for cerebral cortex SA, TH, and subcortical structure volumes were sourced from the ENIGMA consortium⁴⁵. This dataset includes brain cortical thickness and volume data from 51,665 participants across 60 cohorts worldwide, representing diverse ages and genders, with approximately 94% of European ancestry. The cerebral cortex was divided into 34 subregions according to the Desikan-Killiany atlas, and data from corresponding regions in both hemispheres were averaged. Cortical SA measurements were derived from in vivo whole-brain T1-weighted MRI scans, processed using FreeSurfer MRI-processing software. Similarly, subcortical structure volumes, focusing on nine representative subcortical regions, were also sourced from this cohort and processed using the same MRI data analysis methods. Detailed information about these GWAS data sources is provided in Table 1.

5.2.3 Data Sources for smoking and education

Lung function data were partially adjusted for factors related to brain function, such as height and age, but this over-adjustment could introduce bias. Therefore, we utilized MVMR to adjust for variables not accounted for in the GWAS, including smoking and education. Education data were sourced from the Social Science Genetic Association Consortium (SSGAC), encompassing 1,131,881 individuals of European ancestry. The educational attainment was measured based on the age at which participants completed continuous full-time education. Smoking data were derived from the Finnish R10 database, which included 155,619 individuals. Smoking status was defined by the number of cigarettes smoked per day by current and former smokers. Detailed information about these GWAS data sources is presented in Table 1.

5.3 Selection of Instrumental Variables

To select instrumental variables (IVs) from the GWAS summary statistics for lung function metrics, we adhered to the core assumptions of MR and established criteria based on previous literature: (1) Relevance: Genetic variants significantly associated with the exposure traits (FEV1, FVC, and FEV1/FVC) were identified using a p-value threshold of < 5 × 10^-8; (2) Strength: IVs with an F-statistic < 10 were excluded to ensure strong association with the exposure; (3) Linkage Disequilibrium (LD): Potential IVs in LD were excluded using the 1000 Genomes Project data for European ancestry, with a clumping threshold of R^2 = 0.001 within a 10,000 kb distance; (4) Palindromic Variants: After harmonizing the exposure and outcome datasets, palindromic variants were excluded to avoid strand ambiguity; (5) Minor Allele Frequency (MAF): IVs with a MAF < 0.1 were excluded to ensure sufficient allele frequency for robust analysis; (6) Outlier Detection: MR-PRESSO was utilized to identify and remove potential outliers to further ensure the validity of the IVs. After applying these criteria, 260 genetic variants for FEV1, 319 for FVC, and 301 for the FEV1/FVC ratio were identified as suitable IVs for MR analysis. Additionally, for the MVMR analysis, 147 IVs for educational attainment and 152 IVs for smoking behavior met the inclusion criteria. The detailed information regarding these instrumental variables is provided in Supplementary Table S2. These selected IVs were then used in the subsequent MR analyses to examine the causal effects of lung function on brain structure outcomes.

5.4 Mendelian Randomization Analysis

Firstly, we performed univariate MR analyses to explore the effects of FEV1, FVC, and the FEV1/FVC ratio on cortical thickness, cortical volume, and subcortical volume. We utilized five different methods to ensure robust causal inference, with IVW (inverse-variance weighted) as the primary analysis method, supplemented by MR-Egger, weighted median, weighted mode, and simple mode methods. IVW provides the most reliable estimates when all IVs satisfy the three core assumptions of MR. However, when up to 50% of the IVs are invalid, the weighted median method can still offer robust estimates⁴⁶. MR-Egger, although less statistically powerful, can provide supplementary estimates by controlling for potential pleiotropy⁴⁷. The simple mode method remains reliable even when various forms of pleiotropy are present, assuming most of the common pleiotropic effects are zero. Finally, the weighted mode method identifies the largest subset of instruments that provide consistent estimates, thereby enhancing the robustness of our findings⁴⁸. By employing all these methods, we aimed to provide the most comprehensive and reliable causal inference regarding the effects of lung function on brain structure. Each method addresses potential violations of MR assumptions in different ways, ensuring the robustness of our conclusions.

We subsequently conducted a series of sensitivity analyses to confirm the stability of our results. These included heterogeneity testing using Cochran's Q test and pleiotropy analysis using the MR-Egger intercept, MR-PRESSO, and leave-one-out analysis. For Cochran's Q test, a p-value greater than 0.05 indicates no heterogeneity. The MR-Egger intercept is used to estimate horizontal pleiotropy; a p-value less than 0.05 suggests the presence of horizontal pleiotropy among the SNPs, indicating that IVW results may be biased. Leave-one-out analysis systematically excludes each SNP to assess its influence on the overall results, visualized using forest plots to detect potential pleiotropy. For MR-PRESSO, if outliers are detected, they are removed, and the MR analysis is re-conducted. In the MVMR analysis, we adjusted for smoking and educational attainment, reanalyzing the significant results from the univariate MR. The main analysis methods included IVW and MR-Egger, with sensitivity analyses also involving MR-Egger intercept, Cochran's Q test, and leave-one-out analysis, consistent with the univariate MR methods.

All statistical analyses were performed using R software (version 4.2.2), utilizing the “TwoSampleMR” and “MR-PRESSO” packages. All p-values were two-sided. It is important to note that, considering multiple comparisons, we applied the False Discovery Rate (FDR) to correct p-values. A corrected p-value less than 0.05 indicates a causal relationship, while an original p-value less than 0.05 but a corrected p-value greater than 0.05 suggests a potential causal relationship. If both p-values are greater than 0.05, it indicates no statistically significant causal relationship.

Acknowledgments

We thank the researchers and participants of the GWAS studies, including those from the UK Biobank and the ENIGMA consortium, for their invaluable contributions and the data provided for this study.

Author Contributions Statement

Conceptualization, methodology, software, investigation, visualization, and writing of the revised draft: Hongtao You, Chaojuan Huang，Ligang Fan and Xingliang Feng; Data provision and methodology: Chaojuan Huang，Ligang Fan and Xingliang Feng; Conceptualization, writing-revised draft, and funding acquisition: Xingliang Feng, Yuyang Zhang, Naiyuan Shao; Conceptualization, writing-original draft, funding acquisition, and project supervision: Yuyang Zhang, Naiyuan Shao; All authors contributed to the article and approved the submitted version.

Data Availability Statement

All GWAS data used in this study are publicly available. The datasets that support the findings of this study are included in the article and its supplementary materials. Detailed data or code necessary to reproduce our results can be obtained from the corresponding author upon reasonable request.

Ethics Approval Statement

The study was conducted using publicly available GWAS data, which have received ethical approval from the respective study cohorts. No additional ethical approval was required for this study.

Funding

This work was supported by grants from the Nanjing Medical University ChangzhouMedical Center Project (CMCB202312), Changzhou

Municipal Health CommissionMajor Projects (ZD202306) and the Youth science and technology talent lifting project program from Jiangsu and Changzhou Science and Technology Association. The funders had no role in thestudy design, data collection and analysis, decision to publish, or preparation of themanuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Yang, T. et al. Association of fine particulate matter air pollution and its constituents with lung function: The China Pulmonary Health study. Environ. Int. 156, 106707. 10.1016/j.envint.2021.106707 (2021).
Global burden of chronic respiratory diseases and risk factors. 1990–2019: an update from the Global Burden of Disease Study 2019. EClinicalMedicine 59, 101936, doi: (2023). 10.1016/j.eclinm.2023.101936
Christenson, S. A., Smith, B. M., Bafadhel, M. & Putcha, N. Chronic obstructive pulmonary disease. Lancet. 399, 2227–2242. 10.1016/s0140-6736(22)00470-6 (2022).
MacLeod, M. et al. Chronic obstructive pulmonary disease exacerbation fundamentals: Diagnosis, treatment, prevention and disease impact. Respirology. 26, 532–551. 10.1111/resp.14041 (2021).
Marott, J. L., Ingebrigtsen, T. S., Çolak, Y., Vestbo, J. & Lange, P. Trajectory of Preserved Ratio Impaired Spirometry: Natural History and Long-Term Prognosis. Am. J. Respir Crit. Care Med. 204, 910–920. 10.1164/rccm.202102-0517OC (2021).
Wan, E. S. et al. Association Between Preserved Ratio Impaired Spirometry and Clinical Outcomes in US Adults. Jama. 326, 2287–2298. 10.1001/jama.2021.20939 (2021).
Bagdonas, E., Raudoniute, J., Bruzauskaite, I. & Aldonyte, R. Novel aspects of pathogenesis and regeneration mechanisms in COPD. Int. J. Chron. Obstruct Pulmon Dis. 10, 995–1013. 10.2147/copd.S82518 (2015).
Agustí, A., Noell, G., Brugada, J. & Faner, R. Lung function in early adulthood and health in later life: a transgenerational cohort analysis. Lancet Respir Med. 5, 935–945. 10.1016/s2213-2600(17)30434-4 (2017).
Zhou, L. et al. Association of impaired lung function with dementia, and brain magnetic resonance imaging indices: a large population-based longitudinal study. Age Ageing. 51 10.1093/ageing/afac269 (2022).
Ma, Y. H. et al. Lung function and risk of incident dementia: A prospective cohort study of 431,834 individuals. Brain Behav. Immun. 109, 321–330. 10.1016/j.bbi.2023.02.009 (2023).
Shrestha, S. et al. Lung Function and Brain MRI Outcomes in the Atherosclerosis Risk in Communities Neurocognitive Study. J. Alzheimers Dis. 100, 297–308. 10.3233/jad-240162 (2024).
Taki, Y. et al. Correlation between pulmonary function and brain volume in healthy elderly subjects. Neuroradiology. 55, 689–695. 10.1007/s00234-013-1157-6 (2013).
Frenzel, S. et al. Associations of Pulmonary Function with MRI Brain Volumes: A Coordinated Multi-Study Analysis. J. Alzheimers Dis. 90, 1073–1083. 10.3233/jad-220667 (2022).
Yin, M. et al. Patterns of brain structural alteration in COPD with different levels of pulmonary function impairment and its association with cognitive deficits. BMC Pulm Med. 19, 203. 10.1186/s12890-019-0955-y (2019).
Zhang, H. et al. Grey and white matter abnormalities in chronic obstructive pulmonary disease: a case-control study. BMJ Open. 2, e000844. 10.1136/bmjopen-2012-000844 (2012).
Dodd, J. W. et al. Brain structure and function in chronic obstructive pulmonary disease: a multimodal cranial magnetic resonance imaging study. Am. J. Respir Crit. Care Med. 186, 240–245. 10.1164/rccm.201202-0355OC (2012).
Zhu, G. et al. Chickenpox and multiple sclerosis: A Mendelian randomization study. J. Med. Virol. 95, e28315. 10.1002/jmv.28315 (2023).
Yuan, S., Xiong, Y. & Larsson, S. C. An atlas on risk factors for multiple sclerosis: a Mendelian randomization study. J. Neurol. 268, 114–124. 10.1007/s00415-020-10119-8 (2021).
Shrine, N. et al. Multi-ancestry genome-wide association analyses improve resolution of genes and pathways influencing lung function and chronic obstructive pulmonary disease risk. Nat. Genet. 55, 410–422. 10.1038/s41588-023-01314-0 (2023).
Elliott, L. T. et al. Genome-wide association studies of brain imaging phenotypes in UK Biobank. Nature. 562, 210–216. 10.1038/s41586-018-0571-7 (2018).
Smith, S. M. et al. An expanded set of genome-wide association studies of brain imaging phenotypes in UK Biobank. Nat. Neurosci. 24, 737–745. 10.1038/s41593-021-00826-4 (2021).
Dharmage, S. C. et al. Lifetime spirometry patterns of obstruction and restriction, and their risk factors and outcomes: a prospective cohort study. Lancet Respir Med. 11, 273–282. 10.1016/s2213-2600(22)00364-2 (2023).
Higbee, D. H., Granell, R., Sanderson, E., Davey Smith, G. & Dodd, J. W. Lung function and cardiovascular disease: a two-sample Mendelian randomisation study. Eur. Respir J. 58 10.1183/13993003.03196-2020 (2021).
Pai, H. C. & Li, C. C. Relationship Between Trunk Control Ability and Respiratory Function in Stroke Patients: A Scoping Review and Meta-Analysis. Asian Nurs. Res. (Korean Soc. Nurs. Sci). 17, 61–69. 10.1016/j.anr.2023.04.001 (2023).
Mogensen, I. et al. Lung function before and after COVID-19 in young adults: A population-based study. J. Allergy Clin. Immunol. Glob. 1, 37–42. 10.1016/j.jacig.2022.03.001 (2022).
Iversen, K. K. et al. Lung Function Decline in Relation to COVID-19 in the General Population: A Matched Cohort Study With Prepandemic Assessment of Lung Function. J. Infect. Dis. 225, 1308–1316. 10.1093/infdis/jiab636 (2022).
Grenville, J., Granell, R. & Dodd, J. Lung function and cognitive ability in children: a UK birth cohort study. BMJ Open. Respir Res. 10 10.1136/bmjresp-2022-001528 (2023).
Scarlata, S. & Antonelli-Incalzi, R. Poor lung function and associated patterns of cognitive decline. Eur. J. Neurol. 18, 799–800. 10.1111/j.1468-1331.2010.03341.x (2011).
Shrestha, S. et al. Association of Lung Function With Cognitive Decline and Incident Dementia in the Atherosclerosis Risk in Communities Study. Am. J. Epidemiol. 192, 1637–1646. 10.1093/aje/kwad140 (2023).
Wang, J. et al. Pulmonary function is associated with cognitive decline and structural brain differences. Alzheimers Dement. 18, 1335–1344. 10.1002/alz.12479 (2022).
Wang, C. et al. Biomarkers of aging and lung function in the normative aging study. Aging (Albany NY). 12, 11942–11966. 10.18632/aging.103363 (2020).
Wang, C. et al. Altered Gray Matter Volume in Stable Chronic Obstructive Pulmonary Disease with Subclinical Cognitive Impairment: an Exploratory Study. Neurotox. Res. 31, 453–463. 10.1007/s12640-016-9690-9 (2017).
Peng, Z. et al. Cerebral neurovascular alterations in stable chronic obstructive pulmonary disease: a preliminary fMRI study. PeerJ. 10, e14249. 10.7717/peerj.14249 (2022).
Han, Q. et al. Restoring Cellular Energetics Promotes Axonal Regeneration and Functional Recovery after Spinal Cord Injury. Cell. Metab. 31, 623–641e628. 10.1016/j.cmet.2020.02.002 (2020).
Palma, F. R. et al. ROS production by mitochondria: function or dysfunction? Oncogene 43, 295–303, doi: (2024). 10.1038/s41388-023-02907-z
Pregi, N., Belluscio, L. M., Berardino, B. G., Castillo, D. S. & Cánepa, E. T. Oxidative stress-induced CREB upregulation promotes DNA damage repair prior to neuronal cell death protection. Mol. Cell. Biochem. 425, 9–24. 10.1007/s11010-016-2858-z (2017).
Goheen, J., Anderson, J. A. E., Zhang, J. & Northoff, G. From Lung to Brain: Respiration Modulates Neural and Mental Activity. Neurosci. Bull. 39, 1577–1590. 10.1007/s12264-023-01070-5 (2023).
Lim, D. C. & Pack, A. I. Obstructive sleep apnea and cognitive impairment: addressing the blood-brain barrier. Sleep. Med. Rev. 18, 35–48. 10.1016/j.smrv.2012.12.003 (2014).
Li, M. et al. The Protective Effect of Liquiritin in Hypoxia/Reoxygenation-Induced Disruption on Blood Brain Barrier. Front. Pharmacol. 12, 671783. 10.3389/fphar.2021.671783 (2021).
Wang, C. et al. Progress in the mechanism and targeted drug therapy for COPD. Signal. Transduct. Target. Ther. 5, 248. 10.1038/s41392-020-00345-x (2020).
Wilson, D. F. & Matschinsky, F. M. Cerebrovascular Blood Flow Design and Regulation; Vulnerability in Aging Brain. Front. Physiol. 11, 584891. 10.3389/fphys.2020.584891 (2020).
Iadecola, C. et al. The Neurovasculome: Key Roles in Brain Health and Cognitive Impairment: A Scientific Statement From the American Heart Association/American Stroke Association. Stroke 54, e251-e271, doi: (2023). 10.1161/str.0000000000000431
Au Yeung, S. L., Li, A. M., He, B., Kwok, K. O. & Schooling, C. M. Association of smoking, lung function and COPD in COVID-19 risk: a two-step Mendelian randomization study. Addiction. 117, 2027–2036. 10.1111/add.15852 (2022).
Shrine, N. et al. New genetic signals for lung function highlight pathways and chronic obstructive pulmonary disease associations across multiple ancestries. Nat. Genet. 51, 481–493. 10.1038/s41588-018-0321-7 (2019).
Grasby, K. L. et al. The genetic architecture of the human cerebral cortex. Science. 367 10.1126/science.aay6690 (2020).
Bowden, J., Davey Smith, G., Haycock, P. C. & Burgess, S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet. Epidemiol. 40, 304–314. 10.1002/gepi.21965 (2016).
Bowden, J., Davey Smith, G. & Burgess, S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int. J. Epidemiol. 44, 512–525. 10.1093/ije/dyv080 (2015).
Hartwig, F. P., Davey Smith, G. & Bowden, J. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int. J. Epidemiol. 46, 1985–1998. 10.1093/ije/dyx102 (2017).

Table 1. Data sources used in the Mendelian randomization analysis for the current investigation.

Phenotype	Total Sample size	Ancestry	Type of trait	Consortium	PMID
Exposure
FEV1	421,986 individuals	European	Continuous	UK Biobank (IEU GWAS)	221141366
EVC	421,986 individuals	European	Continuous	UK Biobank (IEU GWAS)	221141366
FEV1/FVC	321,047 individuals	European	Continuous	UK Biobank (GWAS Catalog)	30804560
Smoking	155,619 individuals	European	Continuous	Finngen r10	36653562
Education	1,131,881 individuals	European	Continuous	SSGCA	30038396
Outcomes
Total brain cortical SA	51,665 individuals	European (~94%)	Continuous	ENIGMA	32193296
Total brain cortical TH	51,665 individuals	European (~94%)	Continuous	ENIGMA	32193296
SA and TH of 34 regions	51,665 individuals	European (~94%)	Continuous	ENIGMA	32193296

Notes: SD, standard deviation; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; GWAS, Genome-Wide Association Studies; SSGCA, Social Science Genetic Association; SA, surface area; TH, thickness; ENIGMA, the enhancing neuro imaging genetics through meta-analysis.

No competing interests reported.

SupplementaryTable.xlsx

Download PDF

Version 1

posted

You are reading this latest preprint version

Causal effects of lung function on brain cortical structure and subcortical structure: a two-sample univariate and multivariate Mendelian randomization study

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

1 Introduction

2 Results

2.1 Univariate MR results: effects of lung function on cortical structures

2.2 Univariate MR results: effects of lung function on subcortical structures

2.3 Multivariable MR and sensitivity analysis results

3 Discussion

4 Conclusion

5 Materials and methods

5.1 Study Design

5.2 Data Sources

5.2.1 Data Sources for FEV1, FVC, and the FEV1/FVC

5.2.2 Data Sources for cerebral cortex SA, TH, and subcortical structure

5.2.3 Data Sources for smoking and education

5.3 Selection of Instrumental Variables

5.4 Mendelian Randomization Analysis

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1