Причинно-следственная связь метформинового лечения с различными имунизированными подстрекательскими заболеваниями: менделийский анализ случайности.

1. Introduction

immune-mediated inflammatory diseases (IMIDs) include over 100 different diseases, such as inflammatory bowel disease (IBD), rheumatoid arthritis (RA), skin inflammation, and connective tissue disease. [ 1 ] This heterogeneous group of apparently unrelated conditions involving common inflammatory pathways and pathogenic mechanisms. Patients with 1 IMID are at an increased risk of developing another type. [ 2 ] For instance, individuals with RA have a significantly increased risk of developing IBD. [ 3 ] Consequently, the treatment of IMIDs is very important. Metformin is recognized for its ability to reduce inflammation and inhibit autoimmune inflammation, which makes it an attractive exploratory drug.

Metformin is a glucose-lowering agent primarily to treat type 2 diabetes mellitus (T2DM). It has become the first-line medication for individuals newly diagnosed with T2DM. [ 4 ] Metformin also has the effects of the immune-modulating and anti-inflammatory. The effects of anti-inflammatory can operate through both AMPK-dependent and AMPK-independent mechanisms, as demonstrated in various interventional studies using rodent models of obesity and T2DM, as well as in vitro and ex vivo experiments involving several immune cell types. [ 5 , 6 ] Metformin has been reported to exhibit various immunomodulatory effects, including enhancing immunosuppressive capacity in certain autoimmune inflammatory diseases, [ 7 , 8 ] modulating immune responses to infections, and augmenting antitumor immunity. [ 9 , 10 ] Some animal studies also have shown that metformin can treat inflammatory bowel disease, multiple sclerosis (MS) and autoimmune diseases. [ 11 – 13 ] In the collagen antibody-induced arthritis murine model, metformin attenuated arthritis scores, bone destruction, serum levels of the pro-inflammatory cytokines. [ 13 ] In RA, metformin can reduce the joint inflammation and destruction in human cell culture. [ 14 ] And metformin is also effective in the treatment of systemic lupus erythematosus (SLE) and MS. [ 15 , 16 ] However, some studies found no association between metformin treatment and some autoimmune disease. [ 17 , 18 ]

Mendelian randomization (MR) uses genetic variants as instrumental variables (IVs) to evaluate the causal effects of exposure factors on outcomes. This approach can mitigate the influence of confounding factors and reverse causation and the results have been convincing. Therefore, this study employed a 2-sample MR to investigate the relationship investigate metformin and various IMIDs.

2. Methods

2.1. Study design

The workflow of the study is illustrated in Figure 1 . We selected metformin as the exposure and some IMIDs (RA, Crohn disease [CD], ulcerative colitis [UC], SLE, MS, autoimmune hepatitis, primary biliary cholangitis, and primary sclerosing cholangitis [PSC]) as outcomes to conduct a forward MR analysis. The 2-sample MR applied in this study was based on genetic data obtained from publicly available genome-wide association studies (GWAS), which satisfied 3 basic assumptions [ 19 ] : First, the single-nucleotide polymorphisms (SNPs) selected as IVs should have a high degree of correlation with the exposure; Secondly, IVs must be independent of confounders; Finally, IVs should be associated with the outcome solely through metformin and not through alternative pathways.

Flowchart of Mendelian randomization analysis in this study. IMIDs = immune-mediated inflammatory diseases, MR = Mendelian randomization; SNP = single-nucleotide polymorphism.

2.2. Data sources

The exposure data set of metformin was obtained from publicly available GWAS data ( https://gwas.mrcieu.ac.uk/ ). It included 462,933 Europeans individuals (11,552 cases and 451,381 controls) and 9,851,867 SNPs (GWAS ID: ukb-b-14609). [ 20 ] The outcome data were also taken from the IEU Open GWAS project, with the data for RA (GWAS ID: ieu-a-832) [ 21 ] comprised 58,284 samples (14,361 cases and 43,923 controls) and 8,747,963 SNPs, CD (GWAS ID: ukb-a-552) [ 22 ] comprised 337,199 samples (732 cases and 336,467 controls) and 10,894,596 SNPs, UC (GWAS ID: ukb-b-19386) [ 23 ] comprised 463,010 samples (1987 cases and 461,023 controls) and 9,851,867 SNPs, SLE (GWAS ID: ebi-a-GCST90018917) [ 24 ] comprised 482,911 samples (647 cases and 482,264 controls) and 24,198,877 SNPs, MS (GWAS ID: ukb-b-17670) [ 25 ] comprised 462,933 samples (1679 cases and 461,254 controls) and 9,851,867 SNPs, Autoimmune hepatitis (GWAS ID: ebi-a-GCST90018785) [ 24 ] comprised 485,234 samples (821 cases and 484,413 controls) and 24,198,482 SNPs, primary biliary cholangitis (GWAS ID: ebi-a-GCST90061440) [ 26 ] comprised 24,510 samples (8021 cases and 16,489 controls) and 5,004,018 SNPs, and PSC (GWAS ID: ieu-a-1112) [ 27 ] comprised 14,890 samples (2871 cases and 12,019 controls) and 7,891,603 SNPs. The subjects of all datasets are European population. The details of the data source and definition are listed in Table 1 .

Table 1 Characteristics of dataset used for MR analysis. Item GWAS ID Year Case (n) Control (n) Sample size (n) Population SNPs (n) Metformin ukb-b-14609 2018 11,552 451,381 462,933 European 9,851,867  RA ieu-a-832 2014 14,361 43,923 58,284 European 8747,963  CD ukb-a-552 2017 732 336,467 337,199 European 10,894,596  UC ukb-b-19386 2018 1987 461,023 463,010 European 9851,867  SLE ebi-a-GCST90018917 2021 657 482,264 482,911 European 24,198,877  MS ukb-b-17670 2018 1679 461,254 462,933 European 9,851,867  AIH ebi-a-GCST90018785 2021 821 484,413 485,234 European 24,198,482  PBC ebi-a-GCST90061440 2021 8021 16,489 24,510 European 5,004,018  PSC ieu-a-1112 2017 2871 12,019 14,890 European 7,891,603 Abbreviations: AIH = Autoimmune hepatitis, CD = Crohn disease, MR = Mendelian randomization, MS = multiple sclerosis, PBC = Primary biliary cholangitis, PSC = Primary sclerosing cholangitis, RA = rheumatoid arthritis, SLE = Systemic lupus erythematosus, SNPs = single-nucleotide polymorphisms, UC = ulcerative colitis.

Characteristics of dataset used for MR analysis.

Abbreviations: AIH = Autoimmune hepatitis, CD = Crohn disease, MR = Mendelian randomization, MS = multiple sclerosis, PBC = Primary biliary cholangitis, PSC = Primary sclerosing cholangitis, RA = rheumatoid arthritis, SLE = Systemic lupus erythematosus, SNPs = single-nucleotide polymorphisms, UC = ulcerative colitis.

2.3. SNPs selection

The SNPs selected as IVs conformed to the 3 principles described above. And we screened genome-wide for independent SNPs that were strongly associated ( P  < 5 × 10 −8 ) with both the exposure and the outcome. The “TwoSampleMR” R package (version 0.5.8) ( https://github.com/MRCIEU/TwoSampleMR ) [ 28 ] was used to delineate the screening thresholds for linkage disequilibrium ( r 2  < 0.001 and a window size of 10,000 kb) to exclude SNPs exhibiting high linkage disequilibrium. To ensure the consistent effect of IVs alleles across different databases regarding exposure and outcome, harmonization was performed. To eliminate the “weak instruments” ( F -statistic < 10), we calculated the F -statistic for each SNP using the formula F =β 2 /SE 2 . [ 29 ] β represents the estimated genetic effect on exposure, and SE represents standard error of β. Furthermore, to address potential bias due to confounding factors, we visited the PhenoScannerV2 GWAS database ( http://www.phenoscanner.medschl.cam.ac.uk/ ) [ 30 ] and searched for individual phenotypes corresponding to the IVs. We assessed whether these phenotypes were directly associated with the outcome. If they were, the corresponding rs numbers for the phenotypes were identified as confounding SNPs and subsequently excluded. Finally, we remove the SNPs with palindromic properties and incompatible alleles.

2.4. Mendelian randomization analysis

The standard inverse-variance weighted (IVW) method [ 31 ] was used as the primary analysis to calculate the relationship between metformin and IMIDs. Furthermore, the MR-Egger, weighted median, simple mode, and weighted mode methods [ 32 , 33 ] were conducted as supplementary analyses. Currently, IVW is the most widely used statistic in MR analysis, which assume in its simplest form that all genetic variants are effective IVs, providing estimates and precision similar to 2-stage least squares. [ 34 , 35 ]

2.5. Sensitivity analysis

In this study, we performed Cochran Q statistic to test heterogeneity. [ 36 ] The Q -test P  < .05 indicates the presence of heterogeneity, and a random-effects model was used to assess the results. The MR-Egger-intercept test and the outlier (MR-PRESSO) test were utilized to detect pleiotropy. [ 37 ] When the intercept P  > .05, it means that there is an absence of horizontal pleiotropic. Finally, leave-one-out analysis can determine whether any single SNPs are strongly driving causal associations.

2.6. Statistical analysis

All data analyses were conducted in R software (version 4.2.3) with the R packages “TwosampleMR (version 0.5.8).” The difference was considered statistically significant only if the P -value < .05. And the results of MR analyses were expressed as odds ratios (OR) with 95% confidence intervals.

3. Results

3.1. Selection of IVs

IVs associated with metformin (exposure) and IMIDs (outcome) were selected by using R software according to the screening criteria (40 SNPs for RA, 44 SNPs for CD, 45 SNPs for UC, 45 SNPs for SLE, 34 SNPs for MS, 45 SNPs for AIH, 30 SNPs for PBC, 39 SNPs for PSC). Finally, 39 SNPs were selected as IVs for RA due to 1 SNP ( rs11658063 ) being palindromic. And the F -values of these IVs were above the threshold of 10, proving that the screened IVs did not show weak instrument bias.

3.2. Causal effect of metformin treatment on IMIDs

We performed MR analysis on metformin with 8 IMIDs, including RA, CD, UC, SLE, MS, AIH, PBC, and PSC. The IVW analyses were used as the primary method. The results of MR analysis showed that the slopes of the solid lines in the scatter plot of the effects of genetic variants on metformin treatment and IMIDs indicate that the incidence of IMIDs is negatively correlated with metformin treatment. The IVW analysis showed that the treatment of metformin can significantly reduce the risk of RA (OR = 0.018, 95% CI: 1.33 × 10 −3 –0.233, P  = .002), MS (OR = 0.966, 95% CI: 0.936–0.997, P  = .030) and PSC (OR = 6.82 × 10 −4 , 95% CI: 7.83 × 10 −6 –5.93 × 10 −2 , P  = .001). In addition, MR-Egger, weighted median and weighted mode also supported the results of causal associations between metformin and RA except for the simple mode. Between metformin and PSC, of the 4 additional methods carried out, weighted median and weighted mode also suggested significant difference. But for metformin with MS, the other 4 methods all did not support their causal associations. IVW also showed metformin treatment is not significantly associated with the risk of CD (OR = 0.994, 95% CI: 0.979–1.009, P  = .431), UC (OR = 0.987, 95% CI: 0.965–1.009, P  = .234), SLE (OR = 164.373, 95% CI: 0.158–1.71 × 10 5 , P  = .150), AIH (OR = 2.909, 95% CI: 4.58 × 10 −3 –1.85 × 10 3 , P  = .746) and PBC (OR = 0.055, 95% CI: 1.44 × 10 −3 –2.112, P  = .119). Meanwhile, the MR-Egger, weighted mode, simple mode, and WM results were consistent with the IVW. The results of MR analysis as illustrated in Table 2 . Figure 2 illustrates scatter plots depicting metformin to RA, CD, UC, SLE, MS, AIH, PBC, and PSC causal associations. The forest plots are shown in Figure S1, Supplemental Digital Content, http://links.lww.com/MD/O348 . The detailed information of SNPs is listed in Tables S1–S12, Supplemental Digital Content, http://links.lww.com/MD/O347 .

Table 2 The MR analysis results of exposure and outcomes. Outcome Method SNPs (n)
P
OR LCI UCI RA MR-Egger 39 .036 1.14 × 10 −3 2.54 × 10 −6 5.16 × 10 −1 WM 39 .003 3.74 × 10 −3 9.50 × 10 −5 1.47 × 10 −1 IVW 39 .002 1.76 × 10 −2 1.33 × 10 −3 2.33 × 10 −1 Simple model 39 .069 1.49 × 10 −3 1.61 × 10 −6 1.38 Weighted model 39 .003 2.15 × 10 −3 4.55 × 10 −5 1.01 × 10 −1 CD MR-Egger 44 .401 9.83 × 10 −1 9.45 × 10 −1 1.023 WM 44 .201 9.86 × 10 −1 9.64 × 10 −1 1.009 IVW 44 .431 9.94 × 10 −1 9.79 × 10 −1 1.009 Simple model 44 .536 9.87 × 10 −1 9.47 × 10 −1 1.028 Weighted model 44 .336 9.88 × 10 −1 9.64 × 10 −1 1.012 UC MR-Egger 38 .766 1.009 9.53 × 10 −1 1.068 WM 38 .998 9.99 × 10 −1 9.72 × 10 −1 1.029 IVW 38 .234 9.87 × 10 −1 9.65 × 10 −1 1.009 Simple model 38 .153 9.65 × 10 −1 9.20 × 10 −1 1.012 Weighted model 38 .999 1.000 9.68 × 10 −1 1.033 UC (adjusted) IVW 37 .355 9.92 × 10 −1 9.75 × 10 −1 1.009 SLE MR-Egger 45 .468 7.93 × 10 −4 4.03 × 10 −12 1.56 × 10 5 WM 45 .899 1.799 1.90 × 10 −4 1.71 × 10 4 IVW 45 .150 1.64 × 10 2 1.58 × 10 −1 1.71 × 10 5 Simple model 45 .643 1.25 × 10 2 1.90 × 10 −7 8.28 × 10 10 Weighted model 45 .903 2.231 5.46 × 10 −6 9.12 × 10 5 SLE (adjusted) IVW 43 .409 1.44 × 10 1 2.55 × 10 −2 8.13 × 10 3 MS MR-Egger 34 .595 1.021 9.45 × 10 −1 1.104 WM 34 .083 9.77 × 10 −1 9.52 × 10 −1 1.003 IVW 34 .030 9.66 × 10 −1 9.36 × 10 −1 9.97 × 10 −1 Simple model 34 .090 9.51 × 10 −1 8.98 × 10 −1 1.006 Weighted model 34 .198 9.83 × 10 −1 9.57 × 10 −1 1.009 MS (adjusted) IVW 29 8.76 × 10 −8 9.46 × 10 −1 9.27 × 10 −1 9.65 × 10 −1 AIH MR-Egger 45 .646 6.30 × 10 1 1.52 × 10 −6 2.61 × 10 9 WM 45 .765 2.21 × 10 −1 1.13 × 10 −5 4.33 × 10 3 IVW 45 .746 2.909 4.58 × 10 −3 1.85 × 10 3 Simple model 45 .259 4.90 × 10 −5 1.97 × 10 −12 1.22 × 10 3 Weighted model 45 .624 6.41 × 10 −2 1.17 × 10 −6 3.52 × 10 3 PBC MR-Egger 30 .818 3.61 × 10 −1 6.52 × 10 −5 1.99 × 10 3 WM 30 .294 9.76 × 10 −2 1.27 × 10 −3 7.524 IVW 30 .119 5.52 × 10 −2 1.44 × 10 −3 2.112 Simple model 30 .857 2.241 3.70 × 10 −4 1.36 × 10 4 Weighted model 30 .432 1.58 × 10 −1 1.66 × 10 −3 1.49 × 10 1 PBC (adjusted) IVW 27 .132 1.06 × 10 −1 5.73 × 10 −3 1.969 PSC MR-Egger 39 .061 1.87 × 10 −5 3.03 × 10 −10 1.149 WM 39 .008 1.90 × 10 −4 3.548 × 10 −7 1.02 × 10 −1 IVW 39 .001 6.82 × 10 −4 7.834 × 10 −6 5.93 × 10 −2 Simple model 39 .601 2.48 × 10 −2 2.656 × 10 −8 2.31 × 10 4 Weighted model 39 .033 3.38 × 10 −4 2.899 × 10 −7 3.95 × 10 −1 Abbreviations: AIH = Autoimmune hepatitis, CD = Crohn disease, IVW = inverse-variance weighted, LCI = lower confidence interval, MR = Mendelian randomization, MS = multiple sclerosis, OR = odds ratio, PBC = Primary biliary cholangitis, PSC = Primary sclerosing cholangitis, RA = rheumatoid arthritis, SLE = Systemic lupus erythematosus, SNPs = single-nucleotide polymorphisms, UC = ulcerative colitis, UCI = upper confidence interval, WM = weight median.

The MR analysis results of exposure and outcomes.

Abbreviations: AIH = Autoimmune hepatitis, CD = Crohn disease, IVW = inverse-variance weighted, LCI = lower confidence interval, MR = Mendelian randomization, MS = multiple sclerosis, OR = odds ratio, PBC = Primary biliary cholangitis, PSC = Primary sclerosing cholangitis, RA = rheumatoid arthritis, SLE = Systemic lupus erythematosus, SNPs = single-nucleotide polymorphisms, UC = ulcerative colitis, UCI = upper confidence interval, WM = weight median.

The scatter plots of MR analysis. (A) metformin on RA; (B) metformin on CD; (C) metformin on UC; (D) metformin on SLE; (E) metformin on MS; (F) metformin on AIH; (G) metformin on PBC; (H) metformin on PSC. AIH = autoimmune hepatitis, CD = Crohn disease, MR = Mendelian randomization, MS = multiple sclerosis, PBC = primary biliary cholangitis, PSC = primary sclerosing cholangitis, RA = rheumatoid arthritis, SLE = systemic lupus erythematosus, UC = ulcerative colitis.

3.3. Sensitivity analysis

The results of Cochran Q -test for heterogeneity are presented in Table 3 . The analysis revealed significant heterogeneity of metformin to UC ( p IVW  = 0.005), SLE ( p IVW  = 0.035), MS ( p IVW  < 0.001), and PBC ( p IVW  = 0.010). The other groups exhibited no heterogeneity ( P  > .05). Due to the heterogeneity observed in these 4 causality pairs, SNPs with significant differences were excluded through outlier testing. Specifically, 1, 2, 5, and 3 SNPs were removed for the UC, SLE, MS, and PBC groups, respectively. We performed MR analysis again after removing these SNPs with significant differences. The results indicated that metformin treatment significantly reduces the risk of MS (OR = 0.946, 95% CI: 0.927–0.965, P  < .001). The results also suggested that metformin treatment is not significantly associated with the risk of UC (OR = 0.946, 95% CI: 0.927–0.965, P  = .355), SLE (OR = 14.407, 95% CI: 0.026–8.13 × 10 3 , P  = .409) and PBC OR = 0.106, 95% CI: 5.73 × 10 −3 –1.969, P  = .132). The adjusted scatter and forest plots are shown in Figure S2, Supplemental Digital Content, http://links.lww.com/MD/O348 . After adjustment, Cochran Q analysis revealed no significant heterogeneity of metformin to UC ( P  = .610), SLE ( P  = .262), MS ( P  = .776), and PBC ( P  = .574). Meanwhile, in the analysis results of MR-PRESSO, it was found that there was absence of genetic level pleiotropy (Table 3 ). The symmetry of the funnel plot indicated no publication bias (Figure S3, Supplemental Digital Content, http://links.lww.com/MD/O348 ). Additionally, leave-one-out sensitivity testing demonstrated that the causal effect of metformin treatment on IMIDs was not significantly influenced by the exclusion of any single SNP (Figure S4, Supplemental Digital Content, http://links.lww.com/MD/O348 ). These results can be shown to be stable and reliable.

Table 3 Sensitivity analyses of the causal effect of metformin treatment on IMIDs. Outcome Method SNPs (n) Cochran Q -test Pleiotropy test
Q
Q -df
P
Egger-intercept SE
P
RA MR-Egger 39 50.75 37 .066 −9.35 × 10 −3 9.67 × 10 −3 .339 IVW 52.03 38 .064 CD MR-Egger 44 55.99 42 .073 −3.47 × 10 −5 5.89 × 10 −5 .559 IVW 56.46 43 .082 UC MR-Egger 38 61.51 36 .005 −6.75 × 10 −5 8.22 × 10 −5 .417 IVW 62.66 37 .005 UC adjusted MR-Egger 37 32.71 35 .579 −3.58 × 10 −5 6.31 × 10 −5 .574 IVW 33.04 36 .610 SLE MR-Egger 45 59.84 43 .045 3.76 × 10 −2 2.79 × 10 −2 .185 IVW 62.36 44 .035 SLE adjusted MR-Egger 43 45.95 41 .275 2.85 × 10 −2 2.52 × 10 −2 .264 IVW 47.39 42 .262 MS MR-Egger 34 113.93 32 4.05 × 10 −11 −1.70 × 10 −4 1.10 × 10 −4 .134 IVW 122.33 33 3.43 × 10 −12 MS adjusted MR-Egger 29 21.87 27 .744 −7.02 × 10 −5 1.45 × 10 −4 .631 IVW 22.10 28 .776 AIH MR-Egger 45 54.26 43 .117 −9.57 × 10 −3 2.58 × 10 −2 .713 IVW 54.43 44 .135 PBC MR-Egger 30 49.10 28 .008 −6.62 × 10 −3 1.40 × 10 −2 .640 IVW 49.49 29 .010 PBC adjusted MR-Egger 27 24.04 25 .517 −3.25 × 10 −5 1.09 × 10 −2 .998 IVW 24.04 26 .574 PSC MR-Egger 39 50.35 37 .070 1.16 × 10 -2 1.66 × 10 -2 .488 IVW 51.02 38 .077 Abbreviations: AIH = autoimmune hepatitis, CD = Crohn disease, IMIDs = immune-mediated inflammatory diseases, IVW = inverse-variance weighted, MR = Mendelian randomization, MS = multiple sclerosis, PBC = primary biliary cholangitis, PSC = primary sclerosing cholangitis, RA = rheumatoid arthritis, SE = standard error, SLE = systemic lupus erythematosus, SNPs = single-nucleotide polymorphisms, UC = ulcerative colitis.

Sensitivity analyses of the causal effect of metformin treatment on IMIDs.

Abbreviations: AIH = autoimmune hepatitis, CD = Crohn disease, IMIDs = immune-mediated inflammatory diseases, IVW = inverse-variance weighted, MR = Mendelian randomization, MS = multiple sclerosis, PBC = primary biliary cholangitis, PSC = primary sclerosing cholangitis, RA = rheumatoid arthritis, SE = standard error, SLE = systemic lupus erythematosus, SNPs = single-nucleotide polymorphisms, UC = ulcerative colitis.

4. Discussion

Although metformin has been used as a therapeutic agent for T2DM for decades, its molecular mechanisms remain under investigation. The hypoglycemic effects of metformin involve multiple mechanisms. [ 5 ] First, metformin acts on Adenosine 5‘-monophosphate-activated protein kinase (AMPK) to lower hepatic glucose production and reduce peripheral blood glucose concentration. [ 38 ] Second, metformin induces hepatocyte AMPK activation, leading to the downregulation of glycogen gene transcription. [ 39 ] Metformin also acts on complex 1 in mitochondria to promote AMPK activation. [ 40 ] In addition to this, metformin regulates blood glucose levels by inhibiting gluconeogenic enzymes, glucagon signaling, and mitochondrial enzymes. [ 41 – 43 ] In recent years, many studies have found that metformin also has anti-inflammatory and immunomodulatory effects. [ 44 ] Metformin can inhibit the JAK-STAT pathway directly or indirectly via AMPK, limiting interleukin 6 (IL-6)-mediated inflammatory responses. [ 45 ] And the STAT3 signaling pathway is required not only for TH17 cell differentiation but also for T follicular helper (TFH), memory B, plasma cell differentiation and antigen-specific antibodies responses. [ 46 ] Furthermore, AMPK-dependent inhibition of STAT3 and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κBs) pathways are involved in the inhibition of monocyte-to-macrophage differentiation, and reduce the production of inflammatory cytokines by macrophages. [ 47 , 48 ] Therefore, metformin becomes a possibility to treat and prevent autoimmune and inflammatory diseases by regulating multiple cells signaling pathways through AMPK. The abnormal activation of the immune system and inflammatory pathways leads to the systemic involvement of IMIDs.

In this study, we identified a causal effect of metformin treatment on some IMIDs. The results suggested that metformin treatment reduces the risk of RA, MS, and PSC. Although results have indicated that metformin does not significantly decrease the incidence of UC, CD, SLE, AIH, and PBC, it appears that metformin treatment is negatively correlated with the risk of their development, with a tendency to reduce the incidence of these diseases. At present, many articles have shown that metformin treatment can reduce inflammation, delay disease progression and adjust immune system function. But few studies focus on the relationship between metformin and the risk of IMIDs, and these studies are controversial. Naffaa [ 49 ] and Seoyoung [ 50 ] found metformin therapy can reduced the risk of RA in women through Dipeptidyl peptidase-4 (DPP4). However, Zemedikun [ 17 ] and Lu [ 51 ] did not agree with this view. For IBD, there is also no unanimity. In a retrospective study of Taiwan, researchers found the risk of IBD was reduced by half in patients treated with metformin compared to patients treated with other antidiabetic agents. [ 52 ] However, the results of a Danish study showed no association between the history or duration of metformin use and the risk of IBD. [ 53 ] These conflicting results may be related to racial and geographic differences on the 1 hand, and partly due to the bias of retrospective studies on the other. In practice, metformin also may cause side effects, such as acidosis, nausea, abdominal discomfort and diarrhea. [ 54 ] But a randomized trial has confirmed metformin is safe for nondiabetic persons. [ 55 ]

This study is based on a large-scale MR analysis from GWAS summary statistics, circumventing traditional confounding factors and problems associated with reverse causation. Then the F -statistics of all IVs > 10 avoided potential weak instrumental bias. Additionally, several sensitivity tests were conducted to ensure the stability of the results. However, our study does have some limitations. All participants included in the research were of European, which limits the generalizability of our findings. Furthermore, the lack of demographic data in the original study hindered further analyses. Lastly, since the exposure factor in this study was a drug treatment, we were unable to perform bidirectional MR analyses.

5. Conclusion

This study indicated that metformin therapy may reduce the risk of RA, MS, and PSC in European populations, providing new genetic evidence for the relationship between metformin and some IMIDs. Although metformin does not significantly decrease the risk of the other 5 diseases examined, there is a discernible trend indicating that metformin may lower their incidence. The mechanisms by which metformin inhibits IMIDs remain unclear and are a subject of ongoing debate. Further clinical studies and experimental research are necessary to validate the role of metformin in the prevention of IMIDs.

6. Other information

We guarantee that our manuscript is a unique submission and will not be considered for publication in any media by any other source. Furthermore, the manuscript has not been published, in part or in full, in any form. All authors have read and approved the final manuscript and agree to take responsibility for all aspects of the work. All authors have no funding and conflicts of interest to disclose. The datasets generated during and/or analyzed during the current study are publicly available. The summary GWAS data used in this study are publicly available and no specific ethical approval was required. Supplemental Digital Content is available for this article.

Author contributions

Data curation: Zheng Liao.

Project administration: Jinlong Liu.

Software: Zheng Liao.

Supervision: Jian Li.

Validation: Chenguang Su.

Visualization: Chenguang Su.

Writing – original draft: Zheng Liao, Chenguang Su.

Writing – review & editing: Jian Li, Jinlong Liu.

Supplementary Material

Abbreviations: AIH autoimmune hepatitis CD Crohn disease CI confidence intervals GWAS genome-wide association studies IBD inflammatory bowel disease IMIDs immune-mediated inflammatory diseases IVs instrumental variables MR Mendelian randomization MS multiple sclerosis OR odds ratios PBC primary biliary cholangitis PSC primary sclerosing cholangitis RA rheumatoid arthritis SLE systemic lupus erythematosus SNPs single-nucleotide polymorphisms T2DM type 2 diabetes mellitus UC ulcerative colitis

Abbreviations:

autoimmune hepatitis

Crohn disease

confidence intervals

genome-wide association studies

inflammatory bowel disease

immune-mediated inflammatory diseases

instrumental variables

Mendelian randomization

multiple sclerosis

odds ratios

primary biliary cholangitis

primary sclerosing cholangitis

rheumatoid arthritis

systemic lupus erythematosus

single-nucleotide polymorphisms

type 2 diabetes mellitus

ulcerative colitis

Our study was a Mendelian randomization analysis. All data in the article were obtained from genome-wide association study (GWAS) summary data. Therefore, our study did not require ethical approval.

The authors have no conflicts of interest to disclose.

The datasets generated during and/or analyzed during the current study are publicly available.

Supplemental Digital Content is available for this article.

How to cite this article: Liao Z, Su C, Li J, Liu J. Causal association of metformin treatment with diverse immune-mediated inflammatory diseases: A Mendelian randomization analysis. Medicine 2025;104:6(e41400).

ZL and CS contributed to this article equally.

Contributor Information

Zheng Liao, Email: 15511644274@163.com.

Chenguang Su, Email: sucg143143@163.com.

Jian Li, Email: a1291470129@126.com.

References

Associated Data

Supplementary Materials