Fewer Diabetic Complications Seen When NOACs Given for AF

Patients with diabetes and a history of atrial fibrillation (AF) have fewer diabetic complications and lower mortality rates when prescribed non-vitamin K antagonist oral anticoagulants (NOACs) rather than warfarin for stroke prevention, a large nationwide study from Taiwan shows.

Compared with warfarin users, the NOAC group had 16% fewer macrovascular complications, 21% fewer microvascular complications, 9% fewer glycemic emergencies, and 22% lower all-cause mortality at a mean follow-up of 2.9 years, Huei-Kai Huang, MD (National Taiwan University, Taipei), and colleagues report in the study published online February 15, 2022 in Annals of Internal Medicine.

Although reductions in complications and mortality are consistent with what has been seen with NOACs in the diabetic AF cohorts of RCTs, one expert told TCTMD that the drastic declines being reported in this study may be statistical overestimates. And the authors themselves acknowledge that the exact mechanisms by which NOACS may offer more protection against diabetic complications remain unclear, although some prior studies have suggested that NOACs are less likely than vitamin K antagonists such as warfarin to impair insulin sensitivity and glucose tolerance.

“In addition, some preclinical studies have suggested that there are potential effects of direct thrombin or factor Xa inhibitors on anti-inflammatory activities and atherosclerotic plaque stabilization,” the authors write. “The difference in anticoagulation mechanism between NOAC and warfarin could also explain the lower hazard of macrovascular complications with NOAC.”

For the study, Huang and colleagues examined data from the National Health Insurance Research Database of Taiwan, on 19,909 NOAC users and 10,300 warfarin users (mean age 73.8 years; 45.8% female) who had diagnoses of both AF and diabetes. NOAC users were further subdivided by drug: dabigatran (Pradaxa; Boehringer Ingelheim), rivaroxaban (Xarelto; Bayer/Janssen), apixaban (Eliquis; Bristol-Myers Squibb), or edoxaban (Savaysa; Daiichi Sankyo).

To emulate a trial design and strengthen causal inference from the observational data, patients were followed from the first day in 2012 that they began therapy until occurrence of an outcome event, death, or the study’s conclusion on December 31, 2018.

When the researchers analyzed the four NOACs (dabigatran, rivaroxaban, apixaban, and edoxaban) individually, they saw comparable trends toward reduction in diabetic complications and all-cause mortality as the primary analysis. Looking at individual complications, NOAC users had less coronary artery disease, stroke, dialysis, lower extremity amputation, and both CV and non-CV mortality.

Sensitivity analyses confirmed the main results, with lower hazards of macrovascular complications (HR 0.83; 95% CI 0.75-0.92), microvascular complications (HR 0.68; 95% CI 0.61-0.75), glycemic emergency (HR 0.81; 95% CI 0.71-0.93), and all-cause mortality (HR 0.63; 95% CI 0.59-0.68) for NOAC users compared with warfarin users. Propensity score matching of 7,677 pairs showed similarly lower hazards for most outcomes except glycemic emergency, which showed no statistically significant difference between treatment groups.

Inconsistent With RCT Data

Darren McGuire, MD (UT Southwestern Medical Center, Dallas, TX), commenting for TCTMD, said while the study’s conclusion about NOACs versus warfarin is sound and in-line with other reports, including a recent meta-analysis of four phase III randomized controlled trials with nearly 60,000 AF patients, it suffers from many of the pitfalls of observational comparative effectiveness research.

“They've done state-of-the-art methods, but the results are just not believable. There's so much confounding that they can't account for in the propensity adjustments,” he commented. “From randomized trials data of the subsets with diabetes, we get nowhere close to the magnitude of the estimates of efficacy that they are reporting.”

For example, McGuire said, in their propensity analysis, Huang and colleagues report a 50% reduction in CV mortality with NOACs versus warfarin, far more than the 13% that the RCT data show. Similarly, the Huang data show a 44% reduction in stroke with NOACs, while RCT evidence puts it at around 20%.

“Directionally, they have support from randomized trials data, but this study overestimates the efficacy,” he added. “Clearly, NOACs are preferred for these patients, and I don't think anyone would argue that. But we also shouldn't expect a 50% death reduction. There's nothing in the entire world that reduces cardiovascular death by 50%. We have to take these results with a grain of salt.”

As for furthering the understanding of the impact of NOACs on diabetic complications, Huang and colleagues say several studies support the idea that a difference in anticoagulation mechanism between NOACs and warfarin might explain the decreased risk.

“Our study also found that NOAC was significantly associated with a lower dialysis hazard in both the main analysis and the sensitivity analysis, excluding patients with chronic kidney diseases at baseline,” they write. Additionally, while little data exist on retinopathy and neuropathy in diabetic patients receiving oral anticoagulants, their study suggests lower hazards of both with NOACs. “However, future studies are needed to evaluate whether this benefit of NOAC varies between different concurrent antidiabetic medications,” Huang and colleagues say.