BMI and DM: a two-sample mendelian randomization study

doi:10.21203/rs.3.rs-3232845/v1

Download PDF

Article

BMI and DM: a two-sample mendelian randomization study

https://doi.org/10.21203/rs.3.rs-3232845/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Introduction: Diabetes mellitus (DM) patients are often accompanied by an increase in BMI (body mass index). Obesity in clinical work is also often considered to be a high risk factor for diabetes, but whether there is a causal relationship between BMI and DM has not been clearly confirmed. Therefore, we conducted a two sample Mendelian randomization (MR) study to investigate the potential relationship.

Method: We downloaded relevant data of "body mass index", "diabetes" and "type 2 diabetes (T2DM)" from the IEU OpenGWAS project database, used four different methods to conduct two sample MR analysis, and performed sensitivity analysis for verification.

Results: The results of four MR analysis methods indicated that high BMI was indeed an independent risk factor for DM. Since T2DM was a main subtype of diabetes, we further analyzed the impact of BMI on T2DM and got the same results. The two OR values obtained through the IVW method in this study were: OR = 1.032 (95% CI, 1.014 to 1.051), P < 0.001 and OR = 1.005 (95% CI, 1.001 to 1.008), P = 0.011.

Conclusion: Through our research, it was confirmed that high BMI was indeed an independent risk factor affecting the incidence rate of diabetes.

Health sciences/Diseases/Endocrine system and metabolic diseases

Health sciences/Diseases/Endocrine system and metabolic diseases/Diabetes

Health sciences/Diseases/Endocrine system and metabolic diseases/Diabetes/Type 2 diabetes mellitus

Biological sciences/Biological techniques

Health sciences/Diseases

Health sciences/Endocrinology

Health sciences/Health care

Health sciences/Medical research

Health sciences/Pathogenesis

Health sciences/Risk factors

diabetes mellitus

body mass index

Mendelian randomization analysis

risk factors

genome-wide association study

At present, diabetes has become a major chronic disease worldwide, which causes huge burden and harm to the world every year, what's more, it has a trend of becoming younger and younger [1, 2]. At present, diabetes mainly focuses on prevention, supplemented by treatment. Its occurrence and development are related to diet, lifestyle, environment and many other factors. Researchers around the world have begun to use various methods to assess the impact of different factors on the occurrence and development of diabetes, in order to provide relevant evidence for the prevention and treatment of DM. As the main subtype of diabetes, type 2 diabetes has naturally become the focus of lots of researches [3–5]. Since the beginning of the 21st century, over 1 billion people worldwide have become obese, which greatly increases the health burden on all humanity [6]. As we all know, obesity is a poor health condition, which may lead to a series of non communicable diseases, including cardiovascular disease, diabetes and cancer. Related studies found that overweight or obesity may be an independent risk factor for diabetes, especially T2DM [7–9]. In conclusion, it is of great significance to identify high-risk factors related to diabetes for the prevention and treatment of DM. This study aimed to further study the impact of BMI on diabetes.

Mendelian randomization (MR) is a data analysis method that has been widely used in epidemiological etiology inference in recent years. It can strengthen causal inference by using genetic variation as Instrumental variables (IVs). This analysis method is based on Mendelian inheritance law, so the association between genes and diseases can be free from the interference of postpartum environment, socioeconomic status, behavioral factors and other common confounding factors, and the resulting causal sequence is reasonable and closer to the real situation [10]. This research method is conceptually similar to randomized controlled study, in which genetic variations are randomly assigned during gamete formation before being interfered with by any confounding factors, and evenly distributed within the population [11]. Because alleles are fixed between individuals and do not change with the occurrence and development of diseases. Therefore, causal inference got in MR analysis is not easily affected by residual confounding factors and reverse causality [11–13]. In this study, we obtained enough genome-wide association study (GWAS) data from relevant databases, and performed a two sample MR study to assess the impact of BMI on DM based on Mendelian randomization.

2.1. Study design

The premise of Mendelian randomization analysis is that Instrumental variables (IVs) must meet three preconditions: (1) exposure correlation (correlation hypothesis); (2) no common cause with the outcome (independence assumption); (3) outcome related only through the exposure (excluding restriction assumptions). Based on the above criteria, we used MR analysis to explore the causal relationship between BMI and DM [14]. The analysis process of the entire research mainly included five steps: (1) fetching exposure factor GWAS data; (2) sifting appropriate IVs; (3) Inputting the outcome GWAS data and drawing the single nucleotide polymorphism (SNPs) of above IVs; (4) preprocessing the exposure factor and outcome GWAS data to ensure consistency in format; (5) conducting MR and sensitivity analysis.

2.2. Data source

SNPs associated with body mass index were downloaded from the IEU OpenGWAS project database (https://gwas.mrcieu.ac.uk/), using phenotype “BMI” in this study. Its GWAS ID was “ieu-a-2”, the sample size was 339,224 and included 2,555,511 SNPs. The pooled data for diabetes and type 2 diabetes were obtained from the GWAS phenotyped “DM” (GWAS ID:ukb-b-10753; sample size: 461,578; SNPs’ number: 9,851,867) and “T2DM” (GWAS ID: ukb-b-13806; sample size: 462,933; SNPs’ number: 9,851,867), which were also derived from the IEU OpenGWAS project database. The research data was open and transparent, and could be downloaded directly from relevant websites, so no additional ethical declaration or consent was required.

2.3. Selection of IV

We screened out SNPs under the genome-wide significance threshold (P < 5 * 10^-8) related to exposure interest as potential SNPs, visualized the results of the correlation analysis and plotted manhattan plots. In both graphs, the red lines represent the filtering conditions of P < 5 * 10^-8 (Fig. 1). Next, we performed clump function (r^2 = 0.001, kb = 10,000) to eliminate the linkage disequilibrium (LD) between selected SNPs. Further F-statistic was employed to evaluate the effect of weak IVs [15–16], when the F-statistic was less than 10, the genetic variation was considered as weak IV and might cause bias in the research results. After removing the weak IVs, we performed a comprehensive web-based genotype phenotype association database "phenoScanner" to further investigate whether the remaining SNPs were related to potential risk factors of DM, such as long-term bedridden diseases and serious diseases [17–18]. For SNPs associated with confounding factors, we conducted manual screening at the genome-wide significance level (P < 5 * 10^-8) to remove them. After obtaining the remaining SNPs, we made relevant adjustments to make sure that the impact of IVs on exposure and outcome corresponded to the same effector alleles. At last, we removed SNPs with palindrome sequences that could not determine the orientation and incompatible SNPs, and took the remaining SNPs as IVs for MR analysis.

2.4. MR analysis

To further avoid the impact of potential pleiotropy, we employed four different MR methods to assess the causal effect between BMI and DM, namely, inverse variance weighted (IVW), weighted median, simple mode and weighted mode method. Among them, the result of the IVW method was thought as the main result, as the IVW method assumed that all IVs were valid [19–21]. The results of MR analysis were visualized as corresponding plots.

2.5. Sensitivity analysis

To demonstrate the reliability of the research results, we conducted sensitivity analysis to evaluate pleiotropy and heterogeneity. Firstly, the Cochran's Q test was performed to detect potential heterogeneity. When P < 0.05, Cochran's Q statistic evaluated the heterogeneity between genetic variation and the heterogeneity that considered. We visualized the results as corresponding funnel plots. Subsequently, we performed MR-Egger intercept tests to evaluate the horizontal pleiotropy, when P < 0.05, there was pleiotropy in the result. If there was pleiotropy, further analysis and identification of the source for pleiotropy would be conducted through MR-PRESSO analysis. Finally, a leave-one-out analysis was conducted to evaluate whether the causal relationship obtained in the study depended on or leaned towards any single SNP [22].

3.1. Instrumental variables

Through the above screening process, we ultimately screened out 75 SNPs as IVs for MR analysis. The F-statistic of all IVs were > 10, indicating the absence of weak IVs bias.

3.2. MR analysis

The results of all four MR methods used for analysis showed that BMI was positively correlated with the incidence rate of DM. We further investigated the effect of BMI on T2DM and obtained the same results. Specifically, taking the IVW method as the main analysis method, the two OR values obtained in the study were: OR = 1.032 (95% CI, 1.014 to 1.051), P < 0.001 and OR = 1.005 (95% CI, 1.001 to 1.008), P = 0.011 (Fig. 2). Based on the results, we further plotted corresponding scatter plots and forest plots (Fig. 3).

3.3. Sensitivity analysis

Finally, in order to further verify the reliability of the results, we performed sensitivity analysis to examine the heterogeneity and pleiotropy of the conclusion. Cochran's Q test results showed no heterogeneity was found with the IVs included in the study (P > 0.05), and the corresponding funnel plot was shown in Fig. 4. MR-PRESSO analysis did not find significant pleiotropy of the conclusion, nor did it screen out SNPs with outlier (P > 0.05). The test results of the leave-one-out method indicated that the causal relationship between BMI and DM did not depend on or lean towards any single SNP (Fig. 4).

As we all know, diabetes is a chronic metabolic disease caused by many factors. At present, relevant research has found a variety of factors leading to the disease, including genetics, obesity, metabolic syndrome, lifestyle and cardiovascular disease [23–24]. According to the risk factors related to diabetes, researchers have carried out a variety of studies. Many studies have pointed out that the best treatment for diabetes is to prevent the occurrence and development of diabetes by maintaining healthy weight and increasing physical activity [25–27]. Obesity (high BMI) is a poor health condition. Relevant studies have pointed out that it can lead to a variety of non infectious diseases, including cardiovascular diseases and diabetes, etc [7–9]. We have always believed that obesity is a high risk factor for the occurrence and development of diabetes, so in recent years, research on the relationship between obesity and diabetes is in full swing. Yuxia Wei et al. believed that there was a causal relationship between obesity and diabetes, and found that BMI in adulthood was positively correlated with adult latent autoimmune diabetes (LADA) through research [28]. At the same time, their research also pointed out that intervention in childhood obesity was an effective means to prevent diabetes. Another related study also proved that BMI had a direct causal relationship with T2DM, and found that weight control and blood lipid control played an important role in preventing diabetes [29]. The research results of Qiuling Dong et al. clarified the causal relationship between BMI and T2DM from the perspective of metabolic characteristics, and provided new ideas and directions for treatment strategies related to diabetes [30].

Compared with the previous Observational study, this study explored the potential causal relationship between BMI and DM based on four different MR methods. Through MR analysis, we found that high BMI was indeed a high-risk factor for DM, which was consistent with previous clinical experience and the results of numerous studies. The results of sensitivity analysis validation also indicated that the research results we obtained were reliable. The results of four MR methods showed that high BMI increased the risk of diabetes. Based on the analysis results of this study, we could conduct early clinical screening on high-risk (high BMI) populations. In other words, people with high BMI were more likely to suffer from diabetes. We could teach them to avoid diabetes by controlling their weight and strengthening exercise. In addition, obesity has been recommended to be included in the diagnostic criteria of diabetes in some clinical guidelines and academic research, and our research results also provided further theoretical basis for this action.

This study also had some limitations and shortcomings. Firstly, the GWAS dataset obtained in our study came from different populations and years, and further supplementation and improvement could be made in subsequent studies to further avoid the interference of data heterogeneity on the research results. Secondly, diabetes could be divided into many subtypes. We only studied the relationship between BMI and DM, as well as the relationship between T2DM, which accounted for the majority of diabetes. Future research could further explore the relationship between BMI and T1DM. Thirdly, we only excluded SNPs associated with known confounding factors, such as genetics and long-term disease status, future research could be further based on other unknown confounding factors. Finally, we needed to understand that although the “BMI” phenotype could reflect changes in genes, environment, lifestyle and Epigenetics, our research results could only partially explain the causal effect of BMI on DM, and other methods were needed for prove the result in the future.

Through the MR analysis of large sample data, this study proved that high BMI was indeed an independent risk factor for diabetes, and could provide a theoretical basis for the prevention and treatment of diabetes.

Data availability statement

Publicly available datasets were analyzed in this study. This data can be found here: IEU OpenGWAS project: https://gwas.mrcieu.ac.uk/.

Author contributions

Conceptualization: QJ. Methodology: SY and JY. Writing: JY. Review and editing: QJ and SY. All authors contributed to the article and approved the submitted version.

Acknowledgments

We gratefully acknowledge the IEU OpenGWAS project for making their GWAS summary-level statistics publicly available.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Ferguson D, Finck BN (2021) Emerging therapeutic approaches for the treatment of NAFLD and type 2 diabetes mellitus. Nat Rev Endocrinol 17(8):484–495. https://doi.org/10.1038/s41574-021-00507-z
Sun H, Saeedi P, Karuranga S et al (2022) IDF diabetes atlas: global, regional and country-level diabetes prevalence estimates for 2021 and projections for 2045. Diabetes Res Clin Pract 183: 109119. https://doi.org/10.1016/j.diabres.2021.109119
American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2014, 37 (Suppl. S1), S81–S90.
Zheng, Y.; Ley, S.H.; Hu, F.B. Global aetiology and epidemiology of type 2 diabetes mellitus and its complications. Nat. Rev. Endocrinol. 2018, 14, 88–98.
Khan, M.A.B.; Hashim, M.J.; King, J.K.; Govender, R.D.; Mustafa, H.; Al Kaabi, J. Epidemiology of Type 2 Diabetes—Global Burden of Disease and Forecasted Trends. J. Epidemiol. Glob. Health 2020, 10, 107–111.
World Health Organization. Available online: https://www.who.int/news/item/04-03-2022-world-obesity-day-2022-accelerating-action-to-stop-obesity.
Lingvay, I.; Sumithran, P.; Cohen, R.V.; le Roux, C.W. Obesity management as a primary treatment goal for type 2 diabetes: Time to reframe the conversation. Lancet 2022, 399, 394–405.
Lauby-Secretan, B.; Scoccianti, C.; Loomis, D.; Grosse, Y.; Bianchini, F.; Straif, K.; Handbook, I.A.R.C. Body Fatness and Cancer—Viewpoint of the IARC Working Group. N. Engl. J. Med. 2016, 375, 794–798.
Singh, G.M.; Danaei, G.; Farzadfar, F.; Stevens, G.A.; Woodward, M.; Wormser, D.; Kaptoge, S.; Whitlock, G.; Qiao, Q.; Lewington, S.; et al. The Age-Specific Quantitative Effects of Metabolic Risk Factors on Cardiovascular Diseases and Diabetes: A Pooled Analysis. PLoS ONE 2013, 8, e0065174.
Burgess S, Thompson SG (2015) Mendelian randomization: methods for using genetic variants in causal estimation. Chapman and Hall/CRC, London, UK
Smith GD, Ebrahim S. “Mendelian randomization”: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol. (2003) 32:1–22. doi: 10.1093/ije/dyg070
Davey Smith G, Hemani G. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. (2014) 23:R89–98. doi: 10.1093/hmg/ddu328
Burgess S, Butterworth A, Malarstig A, Thompson SG. Use of Mendelian randomisation to assess potential benefit of clinical intervention. BMJ. (2012) 345:e7325. doi: 10.1136/bmj.e7325
Bowden J, Del Greco MF, Minelli C, Smith G, Sheehan N, Thompson J. A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat Med. (2017) 36:1783–802. doi: 10.1002/sim.7221
Burgess, S.; Butterworth, A.; Thompson, S.G. Mendelian Randomization Analysis With Multiple Genetic Variants Using Summarized Data. Genet. Epidemiol. 2013, 37, 658–665.
Zou, X.-L.; Wang, S.; Wang, L.-Y.; Xiao, L.-X.; Yao, T.-X.; Zeng, Y.; Zhang, L. Childhood Obesity and Risk of Stroke: A Mendelian Randomisation Analysis. Front. Genet. 2021, 12, 727475.
Kamat, M.A.; Blackshaw, J.A.; Young, R.; Surendran, P.; Burgess, S.; Danesh, J.; Butterworth, A.S.; Staley, J.R. PhenoScanner V2: An expanded tool for searching human genotype–phenotype associations. Bioinformatics 2019, 35, 4851–4853.
Xu, J.; Zhang, S.; Tian, Y.; Si, H.; Zeng, Y.; Wu, Y.; Liu, Y.; Li, M.; Sun, K.; Wu, L.; et al. Genetic Causal Association between Iron Status and Osteoarthritis: A Two-Sample Mendelian Randomization. Nutrients 2022, 14, 3683.
Burgess S, Bowden J, Fall T, Ingelsson E, Thompson SG. Sensitivity analyses for robust causal inference from mendelian randomization analyses with multiple genetic variants. Epidemiology. 2017;28(1):30–42.
Slob EAW, Groenen PJF, Thurik AR, Rietveld CA. A note on the use of Egger regression in mendelian randomization studies. Int J Epidemiol. 2017;46(6):2094–7.
Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in mendelian randomization with some Invalid Instruments using a weighted median estimator. Genet Epidemiol. 2016;40(4):304–14.
Verbanck M, Chen CY, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. (2018) 50:693–8. doi: 10.1038/s41588-018-0099-7
Bellou V, Belbasis L, Tzoulaki I, et al. Risk factors for type 2 diabetes mellitus: an exposure-wide umbrella review of meta-analyses. PLoS One 2018;13:e0194127.
Fernandes Silva L, Vangipurapu J, Laakso M. The “common soil hypothesis” revisited-risk factors for type 2 diabetes and cardiovascular disease. Metabolites 2021;11:691.
Yuan S, Larsson SC (2020) An atlas on risk factors for type 2 diabetes: a wide-angled Mendelian randomisation study. Diabetologia 63(11):2359–2371. https://doi.org/10.1007/s00125-020-05253-x
Bellou V, Belbasis L, Tzoulaki I, Evangelou E (2018) Risk factors for type 2 diabetes mellitus: an exposure-wide umbrella review of meta-analyses. PLoS One 13(3):e0194127. https://doi.org/10.1371/journal.pone.0194127
Tuomilehto J, Lindström J, Eriksson JG et al (2001) Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 344(18):1343–1350. https://doi.org/10.1056/nejm200105033441801
Wei Y, Zhan Y, Löfvenborg JE, Tuomi T, Carlsson S. Birthweight, BMI in adulthood and latent autoimmune diabetes in adults: a Mendelian randomisation study. Diabetologia. 2022 Sep;65(9):1510-1518. doi: 10.1007/s00125-022-05725-2. Epub 2022 May 23. PMID: 35606578; PMCID: PMC9345833.
Lee Y, Kim YA, Seo JH. Causal Association of Obesity and Dyslipidemia with Type 2 Diabetes: A Two-Sample Mendelian Randomization Study. Genes (Basel). 2022 Dec 19;13(12):2407. doi: 10.3390/genes13122407. PMID: 36553674; PMCID: PMC9777695.
Dong Q, Sidra S, Gieger C, Wang-Sattler R, Rathmann W, Prehn C, Adamski J, Koenig W, Peters A, Grallert H, Sharma S. Metabolic Signatures Elucidate the Effect of Body Mass Index on Type 2 Diabetes. Metabolites. 2023 Feb 3;13(2):227. doi: 10.3390/metabo13020227. PMID: 36837846; PMCID: PMC9965667.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

BMI and DM: a two-sample mendelian randomization study

Status:

Version 1

Abstract

Figures

1. Introduction

2. Method

2.1. Study design

2.2. Data source

2.3. Selection of IV

2.4. MR analysis

2.5. Sensitivity analysis

3. Results

3.1. Instrumental variables

3.2. MR analysis

3.3. Sensitivity analysis

4. Discussion

5. Conclusion

Declarations

References

Additional Declarations

Status:

Version 1