Atrial fibrillation (AF) is the most common cardiac arrhythmia in the clinic, characterized by high rates of thromboembolic complications and related mortality.

An integrated approach of AF care mainly contains rhythm and rate control therapy, anticoagulation therapy (OAC), and comprehensive upstream therapy (1). Among them, OAC is of vital importance in AF-related stroke prevention. As a traditional oral anticoagulant, vitamin K antagonists such as warfarin plays their role by antagonizing vitamin K epoxide reductase complex. Warfarin has a narrow therapeutic range, multiple drug and food interactions, and requires frequent blood monitoring of the international normalized ratio (INR) (2). Therefore, direct oral anticoagulants (DOACs, sometimes referred to as non-vitamin K antagonist anticoagulants) have been introduced in the clinic, reducing the risk of stroke or systemic embolism and bleeding compared with warfarin among patients with AF (3–5).

Patients with comorbid AF and DM have a higher risk of stroke, thromboembolism, and cardiovascular mortality (3, 6–9). Vitamin K has been suggested to regulate the activity of vitamin K-dependent proteins (VKDP) such as osteocalcin, which effectively improve β cell proliferation and insulin secretion to reduce the risk of new-onset DM (9–14). Due to the different effects of warfarin and DOACs on vitamin K, the use of DOACs was considered with the potential of reducing DM risk in AF patients. Two cohort studies have been conducted to compare the risk of DM induction in AF patients treated by warfarin and DOACs (9, 15). The study of Cheung et al. included the data of 13,688 DOACs new users from the Clinical Data Analysis and Reporting System (CDARS) managed by the Hong Kong Hospital Authority (HA). Their results showed that dabigatran was significantly related to incident diabetes risk reduction, and for Xa inhibitor anticoagulants, only the combination use of rivaroxaban and apixaban rather than individual drugs could decrease this risk. In this study, males were the only gender with a diabetic risk-reduction effect. While in a cohort analysis performed by Huang et al., a total of 10,746 AF patients from Taiwan’s National Health Insurance Research Database (NHIRD) were fitted into the study. All of the three different DOACs (dabigatran, rivaroxaban, apixaban) were reported with the incident diabetic reduction effect compared with warfarin. In this study, a similar trend of lowering new-onset DM risk was observed in both male and female groups treated with DOAC vs. warfarin. We used individual participants’ data from the two cohort studies to assess the impact of DOACs on new-onset diabetes risk in AF patients.

The preferred reporting items for systematic review and meta-analysis (PRISMA) 2020 guidelines were used to conduct our present meta-analysis. Only published publications were included in our meta-analysis, so we did not need ethical permission. Readers can contact the corresponding authors for data, techniques, and materials to recreate the results or the program.

The primary findings of this study included (1) Compared with warfarin, DOACs including thrombin inhibitor dabigatran, coagulation factor Xa inhibitor rivaroxaban and apixaban could reduce the risk of new-onset diabetes in AF patients. (2) The DM-reduced effect of DOACs vs. warfarin can be observed in both male and female groups.

Compared with DOACs, warfarin has various limitations in the process of anticoagulant treatment. The changing international standardized ratio (INR) control and dose adjustment, various dietary or drug interactions (16, 17), narrow therapeutic window (17) result in the restrictions of warfarin in clinical use. Several meta-analyses and randomized controlled trials have reported the contrasts with the effectiveness and safety between DOACs and warfarin (18–20). The role of reducing AF and diabetes-associated risk factors including major bleeding, renal decline and cardiac valve calcification can be observed in the use of DOACs rather than warfarin (21–25).

Warfarin plays its role by antagonizing vitamin K, which is an important influence factor of glucose homeostasis and insulin sensitivity. In animal tissues, vitamin K homolog menaquinone-4 (MK-4) might act as an incretin-like nutrient and a cofactor of microsomal γ-glutamyl carboxylase (14, 26, 27). It contributes to the post-translational carboxylation process of transferring glutamate to γ-carboxyglutamate (Gla) residues of VKDP. Insulin production could be promoted by VKDP–osteoblast-specific secreted osteocalcin in a bone-pancreas endocrine loop to regulate glucose metabolism (14). The insulin resistance ameliorating effect of vitamin K was suggested through the inactivation of the NF-κβ signaling pathway to inhibit inflammatory responses and lipid-decreasing effect (11, 12). Considering the vitamin K antagonizing function of warfarin, it can influence the incidence of diabetes. However, the anticoagulation process of DOACs does not influence the vitamin K concentration in the circulatory system. Therefore, using DOACs compared with warfarin could reduce the risk of new-onset diabetes in AF patients.

In 2017, a novel drug betrixaban was approved by FDA as the fifth DOAC that can be used in clinic. With low renal clearance and minimal hepatic metabolism, betrixaban was considered particularly beneficial for patients with renal or hepatic dysfunction (28). However, the lack of an effective reversal agent makes betrixaban has a longer terminal half-life compared with other approved DOACs (28, 29). The impact of betrixaban on the risk of new-onset diabetes has not been evaluated in current studies, and the relevant results are expected to be supplemented in the future.

Our study was based on two cohort studies with a total sample size of 24,434, which is the most comprehensive and latest study according to the risk of DOACs vs. warfarin in inducing new-onset diabetes in AF patients. The result in the cohort study from Cheung et al. (15) proposed that only dabigatran was significantly associated with incident diabetes risk reduction, our results support that three of existing approved DOAC dabigatran, apixaban, and rivaroxaban with the function of reducing incident diabetic risk. At the same time, this effect in factor Xa inhibitors rivaroxaban and apixaban were not obvious, and our findings confirmed that all of these three drugs with the risk reduction ability. Also, in the outcomes of Cheung et al., only a specific gender of AF patients with the advantage of incident diabetic risk reduction in DOACs treatment, whereas the result of our investigation suggested that this effect could be observed in both male and female groups. In addition, the estimated crude events rates of new-onset diabetes were evaluated during our investigation process. Although available data is insufficient to support the effect of vitamin K in ameliorating prediabetes (the impaired glucose tolerance, fasting blood sugar, fasting serum insulin level would not be restored), the glucose and insulin levels of 2-h post-oral glucose tolerance test could be reduced by stable vitamin K support (10, 30, 31), which indicate that DOACs may not induce the rapid deterioration of prediabetes compared with warfarin. Also, the clinical trials have demonstrated that the new anti-diabetic drug sodium-glucose linked transporter inhibitors (SGLTi) with a beneficial effect on cardiovascular disease (32–34), regardless if diabetes exists or not (34). In the aspect of AF, SGLTi can counteract the production of cellular ROS in cardiomyocytes, which may change atrial remodeling and reduce the burden of AF (34). This suggests that the new anti-diabetic drug SGLTi and new oral anticoagulants DOAC may have similar effects on the prevention of new-onset diabetes during AF treatment.

The results of our study suggested that DOACs could reduce the risk of incident diabetes, which is probably more suitable for AF patients with a higher risk of new-onset diabetes. More prospective clinical data about the risk of incident DM in AF patients treated by DOACs and warfarin could prove our point.

This meta-analysis still had several limitations: (1) Only two cohort studies were included in our study, and the data were relatively limited, therefore the evaluation of the effect of edoxaban and betrixaban on incident diabetes was not supported. (2) The included population in our study only contain AF patients from Hong Kong and Taiwan, thus the evaluation of new-onset diabetes risk just considered Asian AF patients. (3) The subtypes of DM were not be distinguished in this study, whether DOACs have the same effect in reducing the risk of type 1 and type 2 diabetes in AF patients is still not clear. (4) The confounding factors cannot be completely excluded in observational studies. According to clinical guidelines, patients with rheumatic heart disease, congenital heart disease, or valve replacement surgery are more likely to be treated with warfarin rather than DOAC (9). This group that may induce selection bias was not completely excluded in our study. Future research could carry out propensity score matching on the basis of incorporating more data to minimize the impact of confounding factors.

Our findings of current analysis suggested that treatment with DOACs compared with warfarin reduced the risk of new-onset diabetes in both male and female patients with AF.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

All authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication.

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.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.