Introduction
Type 1 diabetes mellitus (T1DM) is a chronic autoimmune disease characterized by absolute insulin deficiency and hyperglycemia [
1]. It usually occurs in childhood and adolescence, accounting for approximately 5–10% of all diabetes cases [
2,
3]. There is a growing trend in the prevalence of T1DM globally, with an increase of approximately 2–3% per year. It is predicted that 13.5–17.4 million people worldwide will have T1DM by 2040 (60–107% higher than in 2021), leading to a heavy burden on families and economies [
4,
5]. Pulmonary tuberculosis (PTB) is a chronic infectious disease caused by
Mycobacterium tuberculosis(Mtb) [
6]. According to the World Health Organization (WHO), approximately 10.6 million people suffered from TB in 2021, an increase of 4.5% compared to 2020 [
7,
8]. As the complication rates of diabetes and tuberculosis increase, it is necessary to further elucidate the association to better reduce the burden of disease.
Several recent studies have suggested that T1DM is a primary risk factor for PTB. For example, a hospital-based cross-sectional study reported that the prevalence of Mtb infection in children and adolescents with T1DM was 29.8% (95% CI 24.2–35.4) [
9]. Another case‒control study demonstrated that patients with diabetes had 2.66 times the risk of PTB compared to the general population [
10]. A systematic evaluation of 13 observational studies identified a statistically significant association between diabetes and latent tuberculosis infection [
11]. However, the results of observational studies may be affected by reverse causality and confounding factors, as well as the fact that most of the studies did not specify diabetes phenotypes. Given the differences between T1DM and T2DM in terms of etiology, pathogenesis, and underlying genetic factors, direct evidence of a causal relationship between T1DM and PTB is still currently lacking.
T1DM is frequently accompanied by disorders of glucose and lipid metabolism and obesity [
12,
13]. Some studies have concluded that diabetic patients are susceptible to PTB infection, which is related to disturbed glucose and lipid metabolism in diabetic patients [
14]. A cohort study indicated that an estimated 7.5% of TB occurrences were attributed to poor glycemic control [
15]. Besides, it has been observed that serum high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and total cholesterol (TC) concentrations were lower in patients with PTB [
16]. A systematic evaluation identified a consistent log-linear relationship between BMI and tuberculosis incidence, with a decrease in TB incidence of approximately 14% per unit increase in body mass index (BMI) [
17]. Nevertheless, there are few studies and mostly observational trials. Further research is necessary to support a causal relationship between glucose, lipids, obesity, and PTB.
Mendelian randomization (MR) utilizes genetic variants as instrumental variables (IVs) to elucidate the causal relationship between exposure and outcome [
18]. Due to the random assignment of genetic variants during meiosis, MR is effective in reducing biases caused by confounders and reverse causality [
19]. Compared with prospective experiments, MR analysis reveals causal associations in a time-saving and cost-effective manner [
20].
In the present study, we performed the first bidirectional two-sample MR analysis to assess the causal effects between T1DM and PTB.
Discussion
Based on large-scale GWAS data from MR analyses, our study provided strong evidence that genetic susceptibility to T1DM was associated with an increased risk of PTB, although reverse causality was insufficiently supported. We discovered HDL-C was correlated with the risk of PTB. There was no evidence of an association between other relevant clinical traits of T1DM and PTB.
Our study corroborated that the genetic predisposition to T1DM was associated with an increased risk of PTB, which is consistent with previous studies. Animal studies have indicated an increased susceptibility to Mtb infection in T1DM rats [
41]. Several previous studies have revealed significant heterogeneity in diabetes populations with PTB concerning age, sex, ethnicity, and socioeconomic status, among others [
42]. Nevertheless, it seemed that T1DM remained correlated with PTB after adjusting for potential confounders. A cross-sectional study of children and youth with T1DM indicated that the prevalence of latent tuberculosis was 14.9%, with females slightly higher than males (
P > 0.05), and that the duration of T1DM and age at diagnosis had no significant effect (
P > 0.05) [
43]. Another population-based cohort study showed that patients with T1DM were at an increased risk of tuberculosis, which was higher in men than in women (4.62 vs. 3.59), and in adults than in children (4.06 vs. 3.37), but not significantly [
44]. Nonetheless, these studies were not immune to other confounding factors, and our inability to stratify by gender, age, and other factors was a potential limitation of this study. Moreover, it has been demonstrated that impaired Interleukin-1beta (IL-1β), Interleukin-6 (IL-6), and Interferon gamma (IFN-γ) production in patients with T1DM may lead to increased susceptibility to tuberculosis [
45,
46]. Thus, immune dysfunction may have a significant role in tuberculosis susceptibility. As alveolar macrophages perform a cardinal function in Mtb infection and replication, IFN-γ determines macrophage activation [
47,
48]. IL-1β induces eicosanoids to promote bacterial control and limits type 1 IFN-γ production, which reduces the effect of macrophages and increases tuberculosis susceptibility in patients with T1DM [
49]. Regarding the opposite direction, our MR analyses revealed no evidence to support a causal effect of PTB on T1DM. Despite the studies that have reported a greater risk of diabetes in patients with PTB, evidence is scarce [
50]. Banyai has proposed from animal experiments that Mtb infection can promote necrosis and atrophy of the pancreas to affect diabetes [
51]. Given that T2DM comprises 95% of diabetes, patients in a majority of studies appeared to be T2DM, with the TB-T1DM association currently understudied. Moreover, the results may be attributed to transient hyperglycemia as a result of febrile manifestations of PTB [
52]. A prospective cohort study indicated that PTB could promote transient glycemia, without conclusively demonstrating the promotion of chronic glycemic abnormalities [
53].
The study also explored the correlation between PTB and relevant metabolic characteristics of T1DM. Our purpose was to confirm the influencing factors further by investigating the causal association between glycemic traits and PTB. However, we did not identify a causal correlation between PTB and FBG, HbA1c, and FI, which was different from some previous observational studies [
15,
54]. The difference may be explained as follows: (1) The results obtained from observational studies may be affected by bias, such as confounding factors or reverse causality. (2) Different studies have reached inconsistent conclusions. For example, the cohort study by Pin-Hui Lee et al. indicated that poor glycemic control had a significantly higher hazard of tuberculosis [
15]. Conversely, the opposite conclusion was reported in a Danish population-based case‒control study, which did not demonstrate a significant correlation between glycemic traits and tuberculosis [
55].
In addition, we used several mediators associated with lipid metabolism, including HDL-C, LDL-C, TG, and TC. We found that elevated HDL-C levels increased the risk of tuberculosis. Lipid metabolism, as one of the essential metabolic pathways, can serve as a secondary source of nutrients for PTB infection, favoring growth and multiplication against Mtb. Mtb induces macrophage differentiation into lipid-loaded foam cells and acquires a dormant-like phenotype [
56]. Mtb infection forms granulomas whose core consists of infected macrophages. Progression of granuloma infection is frequently accompanied by dysregulation of lipid metabolism [
57]. It has been suggested that HDL-C enhances Mtb infection in macrophages [
58]. Interestingly, HDL-C plays a dual role in the prevention and regulation of PTB infection. It has been demonstrated that HDL levels are reduced after infection with tuberculosis [
16]. This is probably due to the capacity of HDL to inhibit the production of tumor necrosis factor-alpha (TNF-α), which is critical in the immune defense against TB [
58]. Therefore, further research is still necessary to elucidate the role of HDL-C in different states of TB, such as uninfected, initial infection, asymptomatic state, and active disease.
However, a causal relationship between obesity and tuberculosis was not identified, which differed from several previous studies. A systematic review showed a negative correlation between BMI and the incidence of tuberculosis [
59]. However, results from observational studies may be subject to confounders and reverse causality. For instance, it has been demonstrated that there was a significant correlation between BMI and anemia in patients with PTB, while anaemia as a risk factor for PTB was excluded as a confounder in this study [
60,
61]. Consequently, the MR analysis draws more robust and substantiated conclusions without these issues.
To our knowledge, the study is the first to reveal a causal association between T1DM and PTB using bidirectional MR analysis. Nevertheless, our study has some limitations as well. First, most of the statistics in GWAS were from individuals of European ancestry, which may have raised concerns about the generalizability of the findings to other populations. Ideally, we would repeat this association analysis in large GWAS data from regions with high PTB prevalence (e.g., Africa and South Asia). However, large populations with relevant genomic data are not yet available for further study. Second, despite our efforts to minimize pleiotropy, it was unlikely that all instances of pleiotropy would be eliminated in an MR analysis, which could have biased our results. Third, a potential limitation of the study was the inability to stratify the analyses based on gender, age, and duration of T1DM, among other significant variables.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.