Causal associations between type 1 diabetes mellitus and cardiovascular diseases: a Mendelian randomization study

To investigate the causal relationship between T1DM in a wide range of high-frequency CVD outcomes, we conducted this MR study using large-scale GWAS summary statistics. Our analysis yielded four key findings: (1) genetic predisposition of T1DM was associated with a high risk of both peripheral atherosclerosis and coronary atherosclerosis; (2) the causal effect of T1DM on atherosclerosis is independent of T2DM; (3) hypertension plays an important mediating role in the causal pathway from T1DM to AS; (4) No causal association was observed between T1DM and other CVDs, including HF, AF, CAD, MI, and stroke. Additionally, all the positive outcomes were validated through sensitive analysis (IVW, MR-Egger, MR-PRESSO, leave-one-out analysis, and MVMR).

The process of atherosclerosis always begins at an early stage of life among T1DM patients [8], with endothelial dysfunction being identified as a significant pathophysiological mechanism. Early atherosclerotic changes can be measured by both flow-mediated dilatation (FMD) and carotid intimal medial thickness(cIMT) [36, 37]. A recent meta-analysis focusing on arterial damage in T1DM demonstrated a correlation between high cIMT levels and subclinical arterial injury (mean difference [d] = 0.03, 95% CI = 0.02–0.04) [38]. Mikko J. Järvisalo et al. found impaired FMD was a common manifestation in adolescents with T1DM (4.4 ± 3.4% versus 8.7 ± 3.6%, P < 0.001) [36]. Furthermore, inflammation factors like CRP and IL-6, as well as endothelial markers including sICAM and sVCAM, along with longitudinal lipids, may be associated with higher cIMT and lower FMD, exacerbating endothelial dysfunction and ultimately leading to atherosclerosis [36, 37, 39]. Recent research has identified IL-6 antagonists, such as tocilizumab and ziltivekimab, as potential therapeutic options to improve endothelial function, which can be used as preventive medication for atherosclerosis [40, 41]. Therefore, antagonistic drugs targeting these molecules may potentially attenuate the progression of atherosclerosis in individuals with T1DM. However, further research is needed to explore the efficacy of antagonistic drugs targeting these molecules.

Our findings also highlighted hypertension as a major risk factor for cardiovascular complications among patients with T1DM and its role in promoting the development of atherosclerosis. The potential mechanism underlying hypertension in patients with T1DM may be attributed to diabetic nephropathy [42]. Chronic and persistent hyperglycemia, hyperlipidemia, and glomerular hypertension contribute to the deterioration of renal function [43]. Consequently, there is an accumulation of salts in the body, triggering an excessive activation of the sympathetic nervous system and renin-angiotensin-aldosterone system (RAAS), ultimately leading to hypertension [44]. This process involves various intricate molecules such as transforming growth factor (TGF β1), angiotensin 2(ANG2), vascular endothelial growth factor (VEGF), as well as signal pathways (e.g., TGFβ1-RhoA/Rho signaling) [43]. In clinical practice, inhibition of the RAAS and the use of sodium-glucose cotransporter 2 inhibitors (SGLT2i) have been validated as effective measures to protect kidney from metabolic and hemodynamic damage, which in turn helps to control blood pressure. However, further research focusing on specific cellular pathways from T1DM to diabetic kidney disease is needed and is an area of considerable interest.

In this two-sample bidirectional MR analyses, we did not find any casual relationships between T1DM and HF, MI, CAD, AF, or stroke, which is different from some previous observational studies [9‐11, 45, 46]. A nationwide, register-based cohort study [45] reported that T1DM patients were more susceptible to suffering from acute myocardial infarction [HR = 5.77, 95% CI (4.08–8.16)], stroke[HR = 3.22, 95%CI (2.35–4.42)], as well as heart failure[HR = 5.07, 95% CI (3.55–7.22)]. Maryam Saeed et al. found a nine-fold excess risk of AMI in people with T1DM, [HR = 9.05, 95% CI (7.18–11.41)] [9]. Other cohort studies have reported similar outcomes. This discrepancy can be explained as follows: (1) The outcomes drawn from observational studies are inherently affected by confounding factors. Hence, the impact of T1DM on CVDs may not be as remarkable as previously suggested. (2) Different observational studies have yielded inconsistent conclusions. For example, a population-based prospective cohort study in Sweden did not reveal a significant correlation between T1DM and AF, [HR = 0.99, 95% CI (0.65–1.50)] [10], while another study by Bin Lee Y et al. reported an opposite conclusion[HR = 1.75, 95% CI (1.53 − 1.99)] [11]. To sum up, the results of this MR study suggest that associations between T1DM and CVDs including HF, MI, CAD, AF, and stroke, previously reported in observational studies, may be influenced by biases such as reverse causality or confounding factors.

Extensive research has recently focused on the causal relationship between T2DM and CVD, consistently demonstrating that T2DM is a significant contributor to the development of CVD [15‐17]. In our MR study, we aimed to investigate the causal relationship between T1DM and 7 high frequency CVDs, thereby expanding the knowledge in the field of diabetes and cardiovascular disease research. We specifically emphasized the causal effect of T1DM on atherosclerosis is independent of T2DM. However, our study findings did not identify any significant casual association between T1DM and HF, AF, CAD, MI, or stroke. Some potential mechanisms may partly explain the differences between T1DM and T2DM in the development of CVDs. Although chronic hyperglycemia is a common clinical manifestation and the failure of β cells is a primary event in the development of diabetes mellitus, the nature history, pathophysiologic and genetics mechanisms are all different [47]. Firstly, T1DM typically occurs in adolescence and persists throughout the lifespan, with approximately 80% of pancreatic beta cell function lost by the time of diagnosis. Conversely, T2DM is often diagnosed in middle age when patients still have more than 50% of pancreatic beta cell function [48]. Secondly, autoimmune-mediated β-cell failure leads to absolute deficiency of insulin in T1DM individuals. T cells play a critical role in inducing senescence and apoptosis of pancreatic islet β-cells [5]. On the other hand, there is no convincing evidence supporting autoimmune response in T2DM [49]. T2DM is predominantly mediated by metabolic factors, leading to a sustained decline in β-cell function and insulin resistance in the end [15]. Thirdly, recent research has focused on genetic variants associated with both T1DM and T2DM impacting β-cell [50]. The GWAS have identified more than 400 distinct genetic signals that are evidently associated with T2DM and over 50 signals influencing T1DM [51, 52], highlighting genetic differences between the two types of diabetes. In conclusion, different types of diabetes play distinct roles in cardiovascular complications in spite of the common feature of β-cells failure. More research remains to be done to develop individualized prevention strategies.

The implications of our study for clinical practice are suggested below. Firstly, as T1DM is a lifelong disease, it is significant for physicians to early and regularly evaluate atherosclerotic changes in the arterial wall, especially among adolescents. Secondly, more reasonable strategies for managing hypertension and hyperglycemia need to be developed, as they might slow down the progression of MI, CAD, AF, HF, as well as stroke. Thirdly, in order to ascertain their effectiveness, it is necessary to conduct large-scale randomized controlled trials to validate the potential of novel therapies that aim to protect β-cells in T1DM, including Imatinib [53] and TUDCA [54].

To the best of our knowledge, this is the first MR analysis which aims to find a causal relationship of T1DM on CVDs using the latest and largest GWAS data. We utilized the MR technique to mitigate any potential confounding bias and obtain reliable causal inference. We also employed various crucial methods to systematically investigate the presence of pleiotropy in IVs, which allowed us to address the issue of pleiotropy and enhance the reliability of the MR analysis. Moreover, we observed consistent results across different datasets, which ensures the robustness of the findings.

However, there are still some limitations that need to be emphasized. Firstly, the majority of statistics in the GWAS were derived from individuals of European ancestry, raising concerns about the generalizability of our findings to other populations. Secondly, despite our efforts to minimize pleiotropy, it is unlikely to completely eliminate all instances of pleiotropy in Mendelian randomization studies. There may still be unrecognized pathways and confounding factors between the exposure and outcome variables, potentially introducing biases into our results. Thirdly, a potential limitation of this research is the inability to stratify the analysis based on the severity of T1DM and other important variables such as gender and age.