Assessment of the Relationship Between Genetic Determinants of Obesity, Unhealthy Eating Habits and Chronic Obstructive Pulmonary Disease: A Mendelian Randomisation Study

Abstract

Background: Clinical studies have shown that the onset and exacerbation of chronic obstructive pulmonary disease (COPD) are related to obesity and dietary behaviours, but the genetic relationship between them is not clear.

Aims: To investigate the relationship between the genetic determinants of obesity, dietary habits (alcohol consumption, intake of sweets, salt intake) and COPD.

Methods: Exposure and outcome datasets were obtained from the IEU Open GWAS project. The exposure dataset includes dietary habits (Salt added to food, Sweets intake, Alcohol consumption), obesity level (represented by body mass index (BMI) and body fat percentage (BFP) data sets.). The outcome dataset includes COPD and acute COPD admissions. The collected data were imported into the RStudio software and conducted Mendelian randomisation analysis. Additionally, heterogeneity and horizontal pleiotropy tests were conducted on the data to ensure the veracity of the results.

Results: The results showed that BMI was positively correlated with the risk of acute COPD admission (OR = 1.74, 95% CI 1.39–2.18) and COPD (OR = 1.81, 95%CI 1.41–2.33). In addition, BFP was also a risk factor for COPD (OR = 1.98, 95% CI 1.42–2.77) and acute exacerbation of COPD admission (OR = 1.99, 95%CI 1.43–2.77). The increase of salt, sugar and alcohol consumption will not increase the risk of COPD and the risk of hospitalisation due to COPD.

Conclusion: Therefore, we should strengthen the guidance of diet and living habits of obese patients. For patients with heavier weight and higher body fat rate, they should be instructed to lose weight and fat to prevent the occurrence of COPD. For obese patients with COPD, more attention should be paid to prevent the occurrence of acute exacerbation of COPD in advance.

Keywords:

• chronic obstructive pulmonary disease
• obesity
• mendelian randomisation analysis
• dietary habits

1. Introduction

Chronic obstructive pulmonary disease (COPD) is a lung condition characterised by chronic respiratory symptoms such as shortness of breath, cough and sputum production. It is caused by persistent abnormalities in the airways (bronchitis, bronchiolitis) and/or alveoli (emphysema) and often leads to a progressive decrease in airflow (1). Chronic obstructive pulmonary disease (COPD) is a chronic progressive disease, with respiratory failure being a common cause of death in advanced stages in mild to moderate COPD, cardiovascular diseases and lung cancer are the main causes of death (2). As time goes on, the increasing number of patients with COPD will bring about an increase in the social-economic burden, rising hospitalisation rates, and a decrease in social productivity (3). Early intervention and lifestyle guidance, particularly smoking cessation, have been shown to be the most effective solutions. In addition, recent studies have shown that in addition to smoking, poor eating habits and obesity may also lead to the onset or acute exacerbation of COPD (4–7). Mendelian randomisation (MR) is an analytical method for assessing the genetic causal relationship between observed exposures or risk factors and clinically relevant outcomes (8). In this method, single nucleotide polymorphism (SNP) was used as instrumental variable (IV) to evaluate and analyse outcome, and the correlation between IV and outcome was evaluated by MR-Egger method, the Inverse variance weighted (IVW) method and so on, and then the correlation between exposure and outcome was inferred (9). At present, there is no study using MR method to evaluate the association between obesity, unhealthy lifestyle and the prevalence of COPD and acute COPD. We used MR analysis to further investigate the association between them.

2. Materials and methods

2.1. Data sources

This study used the genome-wide association study (GWAS) to merge the data sets for MR analysis. The selection of the database was based on the release date and sample size. We evaluated the causal relationship between COPD, acute exacerbations of chronic obstructive pulmonary disease (AECOPD) and obesity as well as unhealthy dietary habits, and conducted horizontal pleiotropy analysis and heterogeneity analysis on the data set to validate the reliability of the results. In this study, alcohol consumption, high-sugar diet and high-salt diet intake were classified as poor dietary behaviours, and relevant databases were searched.

Exposure database: The body mass index (BMI) database was obtained from The UK Biobank study and included 461,460 participants. The body fat percentage (BFP) database was from The UK Biobank study, including 454,633 participants. The drinking database was from the Within family GWAS consortium, including 83,626 participants (10). The sugar intake database was from UKB, including 64,949 participants. The salt intake database was also from UKB, including 462,630 participants.

Outcome database: The database of patients with COPD and AECOPD were from the FinnGen alliance. COPD database included 6,915 COPD cases and 186,723 non-cases. The database of acute exacerbations of COPD included 6,500 COPD cases and 212,292 non-cases.

2.2. Instrumental variables

First, to satisfy the assumption of relevance in MR, that is, the IVs must be closely related to the exposure (BMI, BFP, salt added to food, sweets intake, drinking frequency), SNPs significantly associated with the occurrence were selected at the whole-genome level (p < 5 × 10⁻⁸, r² < 0.001, genetic distance= 10,000 KB). If there are too few SNPs included, the inclusion criteria can be changed to (p < 5 × 10⁻⁶, r² < 0.001, genetic distance= 10,000 KB). Additionally, we calculated the F statistics of the IVs to assess the extent of weak instrument bias. To reduce the bias caused by weak working variables, the working variables with F > 10 are retained, and the SNPs with F > 100 are retained for data sets with more SNPs. The formula for calculating the F value and R² is as follows (11).

In order to ensure that genetic variation is not associated with potential confounders, queries were performed in the Phenoscanner database to determine that the included SNPs were not associated with known confounding factors. Finally, after data filtering, horizontal pleiotropy analysis was conducted. If there is horizontal pleiotropy, the MR-presso test is performed. The outliers obtained from the test were removed before proceeding with further analysis.

2.3. MR analysis

MR Egger, weighted median, IVW, simple mode and weighted mode were used for MR analysis to explore the causal relationship between exposure and outcome in this study. Firstly, the MR-presso test was performed on the combined data to screen out the outlier SNPs to remove the level of pleiotropy and to observe whether the screened SNPs met the criteria (p > 0.05). After that, five kinds of analysis are carried out on the data obtained above, and the results of MR analysis are obtained. The positive judgment of the results requires that at least the IVW method has statistically significant results, and the directions of the beta values obtained from the weighted median and MR Egger analyses are consistent (12).

3. Results

3.1. Selection of IVs

The SNP of IV in this study is shown in Table S1. The specific steps are shown in Figure 1. The F statistics for the SNPs included in this study are all greater than 10, indicating that the study results are not biased and confirming the reliability of the results. The impact of each SNP locus on COPD (COPD, acute exacerbation causing hospitalization) was obtained through two-sample MR analysis.

Figure 1. Principle of Mendelian randomization.

3.2. Causal analysis of the relationship between obesity and COPD

In this analysis, there are many SNPs about BMI and BFP, so the IV with F value greater than 100 is selected after calculating the statistical value of F. For this analysis we included to 50 instruments for BMI, 41 instruments for BFP (p < 5 × 10⁻⁸, r² < 0.001, genetic distance = 10,000KB). The IVW method showed that both BMI and BFP were associated with COPD, and the increase in BMI and BFP would increase the risk of COPD. The OR values of the five analysis results are shown in Figure 2. The intercept of the MR-Egger test in the scatter plot was not 0, suggesting that there was pleiotropic effect, but the horizontal pleiotropic test suggested that there was no horizontal pleiotropic effect, indicating that the pleiotropic effect of exposure had no effect on the outcome. The level pleiotropy test showed that there was no level pleiotropy. The results of leave-one-out analysis showed that some SNPs exceeded the invalid vertical line. However, the occurrence of this situation does not have a negative impact on causal inference. Scatter plot, leave-one-out sensitivity analysis plot, forest plot and funnel plot are shown in Supplementary Table 1.

Figure 2. The five analysis results of MR* analysis between obesity and COPD*.

*MR: Mendelian randomization; COPD: chronic obstructive pulmonary disease

3.3. Causal analysis of the relationship between obesity and AECOPD

In this analysis, there are many SNPs about BMI and BFP, so the IV with F value greater than 100 is selected after calculating the statistical value of F. For this analysis we included to 50 instruments for BMI, 41 instruments for BFP (p < 5 × 10⁻⁸, r²<0.001, genetic distance = 10,000KB). The IVW method showed that BMI (MR-PRESSO test showed that there was outlier SNP, and the pvalue was 0.7498 after removal) and BFP was associated with admission to acute exacerbation of COPD and was a risk factor for admission to acute exacerbation of COPD. The OR values of the five analysis results are shown in Figure 3. The intercept of the MR-Egger test in the scatter plot was not 0, suggesting that there was pleiotropic effect, but the horizontal pleiotropic test suggested that there was no horizontal pleiotropic effect, indicating that the pleiotropic effect of exposure had no effect on the outcome. The results of leave-one-out analysis showed that some SNPs exceeded the invalid vertical line. However, the occurrence of this situation does not have a negative impact on causal inference. Scatter plot, leave-one-out sensitivity analysis plot, forest plot and funnel plot are shown in Supplementary Table 2.

Figure 3. five analysis results of MR analysis between obesity and AECOPD*.

*AECOPD: Acute exacerbations of chronic obstructive pulmonary disease

3.4. Causal analysis of the relationship between dietary habits and COPD

For this analysis, we included to 106 instruments for salt intake, 84 instruments for alcohol consumption, 21 instruments for sweets intake (less SNPs for sweets intake during screening, pvalue was changed to 5 × 10-6 for screening). IVW method showed that all three exposures were not related to the incidence of COPD (MR analysis results are shown in Figure 4). OR values were salt intake (OR = 1.12,95%CI, 0.79-1.59, p = 0.54), alcohol consumption (OR = 0.94,95% CI, 0.67-1.31, p = 0.72), sweets intake (OR = 0.89,95% CI, 0.64-1.2, p = 0.53). Scatter plot, leave-one-out sensitivity analysis plot, forest plot and funnel plot are shown in Supplementary Table 3.

Figure 4. MR analysis results between dietary habits* and COPD.

*Dietary habits: This includes alcohol consumption, intake of sweets, salt intake.

3.5. Causal analysis of the relationship between dietary habits and AECOPD

For this analysis, we included 106 instruments for salt intake, 84 instruments for alcohol consumption, 21 instruments for sweets intake (less SNPs for sweets intake during screening, pvalue was changed to 5 × 10-6 for screening). IVW method showed that all three exposures were not related to the incidence of AECOPD (MR analysis results are shown in Figure 5). OR values were salt intake (OR = 1.0795%CI 0.742–1.53, p = 0.73), alcohol consumption (OR = 1.02, 95%CI 0.95–1.09, p = 0.62), sweets intake (OR = 0.94, 95%CI 0.66-1.33, p = 0.71). Scatter plot, leave-one-out sensitivity analysis plot, forest plot and funnel plot are shown in Supplementary Table 4.

Figure 5. MR analysis results between dietary habits and AECOPD.

4. Discussion

COPD is a heterogeneous lung disease, which is often caused by abnormalities in the airways (bronchitis, bronchiolitis) and/or alveoli (emphysema). Patients with COPD show chronic respiratory symptoms such as dyspnoea, cough, and sputum caused by progressive airflow obstruction (13–16). With the continuous ageing of the world, the number of patients with COPD and the increasing number of hospitalisations due to increased COPD have caused a huge burden on society (17, 18). Furthermore, the number of deaths due to COPD is also increasing, and COPD is expected to be the fourth leading cause of years of life lost (YLLs) by 2040 (19). Apart from those who have already been diagnosed, there are still many undiagnosed COPD patients. Compared with healthy people, the respiratory symptom burden of undiagnosed COPD patients has a greater impact on their health and daily activities, leading to poor quality of life, overall poorer health, increased utilisation of medical services and negative effects on work productivity (20).

Smoking is considered the main cause of this disease, while dust, smoke, air pollution and exposure to harmful gases related to the working environment are also gradually attracting attention as causes of the disease (21–24). In addition, the correlation between obesity and COPD has long been recognised, but there has been controversy over whether obesity is a risk factor for COPD. Many studies have shown that overweight COPD patients have higher survival rates and better lung function performance compared to COPD patients with normal weight (25–31). Although there is data showing that obesity is more common in the COPD population than in the general population, few studies have shown whether the incidence of COPD is different between obese patients and the general population (26, 32). Generally, BMI is widely recognised as an indicator for measuring obesity, but BMI cannot distinguish whether overweight is mainly due to fat or muscle mass (33). Zhang and his team analysed data from the National Health and Nutrition Examination Survey (NHANES) from 2007 to 2012 and found that underweight and abdominal obesity are associated with an increased risk of airflow obstruction, while overweight and general obesity are associated with a decreased risk of airflow obstruction, reflecting the existence of the ‘obesity paradox’ (34). Lavie and other researchers further explored this conclusion and proposed that the presence of physical activity (PA) and cardiorespiratory fitness may affect the outcomes of overweight patients with COPD, and these two factors are the main reasons for the existence of the ‘obesity paradox’ (35). Therefore, this study used BFP as a supplementary variable for analysis. The results of this study showed that patients with high BMI and BFP were more likely to suffer from COPD. This explains to some extent the reason for the existence of the ‘obesity paradox’, that is, patients with high BMI may not have high BFP, so these patients with high BMI have better lung function test results. Although these mechanisms need to be revealed by further studies, this finding is also worth noting. Compared with observational studies, MR analysis has obvious advantages, including the ability to avoid bias caused by confounding factors and reverse causality (8). In terms of the hierarchy of evidence, MR is between interventional epidemiology and observational epidemiology (36). Previous studies have suggested that COPD patients with obesity have a higher admission rate and are more prone to acute exacerbation (37–39). Larsson mentioned in a literature about the causal relationship between high BMI and various chronic diseases that excessive obesity is associated with an increased risk of COPD, which is consistent with the results of this article (40). The results of this study are consistent with previous studies that both BMI and BFP increase the risk of acute exacerbation of COPD.

This study used MR method to evaluate the relationship between alcohol consumption, salt intake, sugar intake and COPD. The study concluded that the three were unrelated to the occurrence and AECOPD. A study enrolled 16,907 participants to analyse the incidence of COPD and consumption of sugar-sweetened soft drinks. The results showed that consumption of sugar-sweetened soft drinks was positively correlated with COPD, and it was inferred that excessive sugar intake may contribute to airway inflammation, although the underlying mechanism remains unclear (41). This situation may have occurred because the population involved in this experiment was Australian, which has a different genetic background from the European ethnic groups included in this study. A prospective study followed 30,503 patients for a median of 3.35 years to investigate the relationship between drinking and AECOPD, and the results showed that drinking was unrelated to AECOPD (42). This study used MR method to evaluate the relationship between alcohol consumption, salt intake, sugar intake and COPD. The study concluded that the three were unrelated to the occurrence and AECOPD. Notably, there is currently no cohort study exploring the potential relationship between salt intake and the incidence and acute exacerbation of COPD.

MR analysis primarily relies on data from GWAS databases, which enhances its reliability compared to conventional cohort studies. To bolster the credibility of the research findings, this study employed five MR methods, along with heterogeneity analysis and horizontal pleiotropic analysis. Nonetheless, there are certain limitations to consider: The GWAS data included in this study represent the European population, therefore caution should be paid when inferring causal relationships for other geographic origins of data.

Conclusion

This study found that obesity increased the risk of COPD and acute exacerbation of COPD by MR analysis, and drinking, sugar intake and salt intake were not associated with COPD and acute exacerbation. Therefore, we should strengthen the guidance of diet and living habits of obese patients. For patients with heavier weight and higher body fat rate, they should be instructed to lose weight and fat to prevent the occurrence of COPD. For obese patients with COPD, more attention should be paid to prevent the occurrence of acute exacerbation of COPD in advance. In addition, due to the genetic differences between different geographic origins of data, different populations need to be further studied.

Author contributions

Tongyao Sun: conceptualisation, methodology, formal analysis, data curation, writing-original draft, writing-review and editing. Jun Wang: data curation, writing-review and editing, visualisation. Chengsen Cai, Jianjian Yu, Lina Fu, Lei Duan: data curation, writing-review and editing, visualisation. Min Zheng: conceptualisation, methodology, formal analysis, supervision, writing-original draft, writing-review and editing. All authors read and approved the final manuscript.

Ethical Statement

Ethics approval was not needed as this study did collect any primary data from human participants or animals.

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.

Disclosure statement of competing interest

The authors declare that they do not have any competing interests.

Additional information

Funding

Supported by Shandong Province Pharmaceutical Health Science and Technology Development Plan Project (202003020981) and Health Commission of Shandong Province (2021Q101).