Medical Therapies for Prevention of Cardiovascular and Renal Events in Patients With Atrial Fibrillation and Diabetes Mellitus

Laurent Fauchier; Giuseppe Boriani; Joris R. deGroot; Reinhold Kreutz; Peter Rossing; A. John Camm


Europace. 2021;23(12):1873-1891. 

In This Article

Beyond Macrovascular Complications of Atrial Fibrillation With Diabetes: Heart Failure and Chronic Kidney Disease

Mechanisms, Risks, and Optimal Management of Patients who Have or are at Risk of Developing Heart Failure

The first letter 'C' representing congestive HF in the CHA2DS2-VASc score, highlights the predictive role of HF for stroke risk in AF which is particularly important in diabetic patients who are at increased risk for both stroke and HF.[28,41] The prevalence of DM in HF patients is about 30–40% and similar in HF with reserved ejection fraction (HFpEF) and HF with reduced ejection fraction (HFrEF).[28,111,112] Patients with DM are at higher risk of developing HFrEF or HFpEF[28] and the risk increases with age.[113–115] Conversely, HF itself is a risk factor for the development of DM,[116] most likely related to insulin resistance.[28] The presence of HF results in a higher risk of HF hospitalization, CV death, and all-cause death in patients with DM,[6,28] with the strongest predictive value of DM seen in patients with HFrEF.[28] At the time of new-onset AF, patients with DM are at increased risk for developing acute HF due to the loss of atrial kick and impaired LV filling.[117] AF and acute HF frequently co-exist and can exacerbate each other.[118] In acute HF, the risk is magnified by the presence of DM, because the DM will significantly increase short-term risk including in-hospital death and risk for rehospitalization due to HF and all-cause death within 1 year.[117,119,120]

In addition to the direct detrimental effect of insulin resistance and hyperglycaemia on LV dysfunction, major risk factors promoting the development of HF in DM are CAD, CKD, and hypertension.[28] LV diastolic dysfunction is frequent in DM and already observed in patients with pre-diabetes as it shows a strong correlation with insulin resistance and hyperglycaemia.[28,121] Heart failure with reserved ejection fraction represents the most frequent form of HF in DM (about 75%) and its prevalence is higher in older, female, and hypertensive patients with DM.[122] Guideline-directed medical treatment and device therapies for HF are equally effective in patients with and without DM as shown in RCTs in which on average about 30–40% of patients had DM.[28]

Beta-blockers play per se an important role for rate and symptom control in patients with AF.[41] They are the main part in the 'B' pillar of the 'ABC' AF management ('B': better symptom control with rate or rhythm control therapies)[41] and fully relevant for rate control in AF patients with DM. Interestingly, while their treatment benefits strongly support their general use also in patients with HFrEF and DM in sinus rhythm,[28,41] their prognostic benefit in patients with AF has been questioned.[123] In patients with acute HF, beta-blockers should be cautiously initiated in the hospital, once the patient is stabilized.[124]

Non-dihydropyridine calcium channel blocker (verapamil or diltiazem) can be used as an alternative to beta-blockers for rate control in AF patients and for treatment of hypertension in patients with or without HFpEF, but their use is contraindicated in patients with HFrEF.[41,124] In addition, due to their inhibitory effect on of P-glycoprotein and the cytochrome P 450 3A4 enzyme, drug interactions between verapamil and diltiazem with DOACs should be considered. They could lead to higher drug levels of DOACs and thus increased bleeding risk.[41,125,126] Among patients with permanent AF and symptoms of HF, low-dose digoxin may also be used considering that this may result in a similar quality of life at 6-months than with bisoprolol.[127]

For several decades, the available evidence from clinical studies indicated that glycaemic control in DM with glucose-lowering therapies had, if anything, only moderate beneficial effects on macrovascular endpoints and on the risk for HF-related outcomes.[128] However, the clinical development of SGLT2 inhibitors and glucagon-like peptide 1 receptor agonists during recent years, with the completion of several important CV outcome trials resulted in fundamental changes.[28,74]

Particularly, the SGLT2 inhibitors have been shown to significantly lower the risk for hospital admissions for HF in T2DM patients at high CV risk.[70–72,129,130] The risk reductions for HF hospital admission were consistent and in the range of 27–39% in these trials. However, the prevalence of patients with AF in these studies was relatively low or originally not reported.[70–72,129,130] Nevertheless, a dedicated post hoc study compared patients with AF at baseline (n = 389) and patients without AF (n = 6631) in the EMPA-REG OUTCOME trial.[131] This analysis demonstrated that empagliflozin compared to placebo reduced CV death or HF hospitalization consistently in diabetic patients with AF (HR 0.58, 95% CI 0.36–0.92) and without AF (HR: 0.67, 95% CI 0.55 to −0.82, P-int = 0.56).[131] Of interest, the absolute number of prevented events was higher in patients with AF, resulting in larger absolute treatment effects of empagliflozin.

A subgroup analysis investigating the impact of AF (about 38% of the study population) is available in DAPA-HF.[132] It showed in patients with AF at baseline a somewhat weaker protective effect (HR 0.82; 95% CI 0.63–1.06) compared to patients without AF (HR 0.72; 95% CI 0.61–0.84) on the primary composite outcome of worsening HF or CV death.[132] This analysis included, however, both HFrEF patients with and without DM as enrolled in DAPA-HF. Thus, more dedicated analyses on the effects of SGTL2 inhibitors in diabetic patients with AF are needed. More recently, however, a study investigating the effect of the combined SGLT2 and SGLT1 inhibitor sotagliflozin in patients with DM and a history of recent worsening of HF, confirmed the protective effect of this treatment approach.[133] In this trial, a subgroup analysis showed in patients with AF/atrial flutter (n = 576, HR 0.68, 95% CI 0.48–0.95) a similar risk reduction as compared to patients without AF/atrial flutter (n = 646, HR 0.69, 95% CI 0.49–0.97) for the primary composite outcome of the total number of deaths from CV causes and hospitalizations and urgent visits for HF.[133]

Mechanisms, Risks, and Optimal Management of Patients who Have or are at Risk of Developing Chronic Kidney Disease

Diabetic Kidney Disease. In addition to being a risk factor for the development of AF and for complicated stroke in the presence of AF, DM is associated with an increased risk for the development of kidney disease. On average 30–40% of patients with DM develop diabetic kidney disease or CKD in DM, with increasing albuminuria and declining renal function as well as increased risk for CV disease.[32] Diabetes mellitus is the leading cause of kidney failure in most parts of the world accounting for up to 50% of kidney failure, but the majority of patients with DM and CKD die from CV events before reaching kidney failure.

There seems to be a familial disposition for diabetic kidney disease, and many genetic variants have been associated with diabetic kidney disease, although major single-gene effects have not been demonstrated.[134,135] In addition to hyperglycaemia, risk factors for CKD in DM include hypertension and smoking (Figure 8). Markers of risk also include oxidative stress, endothelial dysfunction, inflammation, uric acid, and dyslipidaemia.[136]

Figure 8.

Factors potentially responsible for the high prevalence of chronic kidney disease in diabetic patients.

Comprehensive management of CKD in DM includes lifestyle factors such as exercise, a healthy diet with a focus on protein intake and salt, and smoking cessation is important. For the kidney as well as the CV risk, management of blood pressure with blockade of the renin-angiotensin-system (RAS), including use of angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs), as well as lipid control and glucose management is important.[137]

In T2DM, SGLT2 inhibitors are recommended for their protective effect on kidney function.[138] They also have beneficial effects on CV disease and HF as seen in the dedicated kidney studies CREDENCE (Evaluation of the Effects of Canagliflozin on Renal and Cardiovascular Outcomes in Participants With Diabetic Nephropathy) with canagliflozin[72] and DAPA-CKD with dapagliflozin,[139] including patients with urinary albumin creatinine ratio from 200 mg/g creatinine and eGFR from 25 to 90 mL/min/1.73 m2.

As mentioned in Mechanisms, risks, and optimal management of patients who have or are at risk of developing heart failure section, in a post hoc analysis the SGLT2 inhibitor dapagliflozin reduced the incidence of AF in the DECLARE TIMI 58 study, which was a CV outcome study in T2DM with previous CVD or risk factors for CVD.[73]

Anticoagulation-related Nephropathy. In patients with AF and CKD treated with anticoagulants, there is an increased risk of bleeding, but there are also concerns about the risk of anticoagulation-related nephropathy (Figure 9).[140] The latter is a newly recognized form of acute kidney injury in which over-anticoagulation causes profuse glomerular haemorrhage, which manifests on renal biopsy as numerous renal tubules filled with red cells and red cell casts. The glomeruli show changes, but they are not sufficient to account for the glomerular haemorrhage.[140]

Figure 9.

Possible mechanisms for changes in renal function in patients treated with anticoagulants. PAR, protease-activated receptors. aDose dependent/or during low sustained activation.

Although anticoagulation-related nephropathy (Figure 9) may develop in response to any anticoagulant including VKAs and DOACs, it has been particularly associated with overdosing of warfarin with International Normalized Ratio (INR) levels >3.[140] Older patients with DM and diabetic kidney disease are particularly prone to develop anticoagulation related nephropathy that may trigger episodes of acute kidney injury more frequently in AF patients than previously thought.[140,141] These acute kidney injury episodes can in turn accelerate the progression of CKD and are associated with an increased mortality rate.[142]

Both DM and CKD increase the risk of stroke and other CV complications in patients with AF and as discussed in the other sections DOACs are preferred to VKA. In patients with significantly impaired renal function (CrCl < 30 mL/min) a reduced dose of rivaroxaban, apixaban, or edoxaban is recommended in patients with a CrCl between 15–30 mL/min, but patients with a CrCl <25–30 mL/min were not included in the randomized trials. The FDA approved a low 75 mg bid dose of dabigatran for patients with a CrCl <30 mL/min, available in the USA, but it has not been tested in a prospective trial, and it is not approved for use within Europe. Vitamin K antagonist has been suggested if time in therapeutic range (TTR) > 70% but harm may exceed benefit with increased risk for bleeding, anticoagulation-nephropathy, and vascular calcification. Until ongoing studies clarify the optimal anticoagulation strategy in severe kidney disease, this has to be based on individual assessment of risk and benefits.

Vascular Calcification Theory. Diabetes mellitus and CKD share complementary pathophysiology for increased vascular calcification processes[143,144] which may contribute to declining renal function in patients with DM and CKD. In addition to glucose-related pathways in DM, further factors related to CKD such as disturbance in calcium–phosphate balance, accumulation of uraemic toxins, and severe vitamin K deficiency have been implicated in the pathogenesis of vascular calcification. In anticoagulated AF patients and particularly in patients with comorbid diabetic kidney disease, the use of VKAs may aggravate vascular calcification[145] including calcification in the vascular bed of the kidney and thereby contribute to worsening renal function in these patients (Figure 9).[146,147] This is mechanistically based on inhibition of the vitamin K-dependent gamma-glutamyl carboxylation that applies not only to the clotting factors II, VII, IX, and X but also to the other vitamin K dependent gamma-carboxyglutamic acid (Gla) proteins including the matrix Gla protein.[144,148] Matrix Gla protein represents the most potent endogenous protector against vascular calcification and its diminished function has been linked to vascular calcification during VKA treatment.[148] In contrast, DOACs, such as the factor Xa inhibitor rivaroxaban, due to their different mode of action do not only lack this negative effect but may even provide beneficial protective effects against vascular injury and renal functional decline by decreasing vascular inflammation, remodelling, and vascular calcifications through reduced protease-activated receptor (PAR) signalling via PAR-1 and PAR-2 (Figure 9).[149,150]

In accordance with this, observational studies of subjects with AF treated with VKAs vs. DOACs found more progression of kidney disease (development of CKD Stage 5, i.e., eGFR < 15 mL/min/1.73 m2 or need for kidney replacement therapy) with VKAs than DOACs. Thus, a recent study using USA IBM MarketScan data included patients with AF and DM that were newly initiated on rivaroxaban (N = 10 017) or warfarin (N = 11 665).[106] Patients were matched using propensity scores. In comparison to warfarin, rivaroxaban was associated with lower risks of acute kidney injury events (HR 0.83, 95% CI 0.74–0.92) and development of Stage 5 CKD or need for haemodialysis (HR 0.82, 95% CI 0.70–0.96). The protective effect in favour of rivaroxaban was particularly pronounced in the subgroup of diabetic patients with pre-existing Stages 3–4 CKD for both acute kidney injury (HR 0.63, 95% CI 0.49–0.79) and the risk of stage 5 CKD or need for haemodialysis (HR 0.66, 95% CI 0.46–0.94).[106] A similar retrospective study on data from a claims database in Germany found the relative risks for acute kidney injury was decreased by 28% (HR 0.72, 95% CI 0.53–0.97) and for kidney failure by 68% (HR 0.32, 95% CI 0.19–0.53) in AF patients with DM prescribed rivaroxaban vs. phenprocoumon.[151] These observational findings should be confirmed in prospective randomized controlled trials.

Kidney Function Monitoring. Due to the high risk for kidney disease, screening for this complication is mandated in the regular Follow-up of all patients with DM, with annual measurements of urine albumin excretion (urinary albumin to creatinine ratio in morning spot urine) and assessment of renal function (estimated glomerular filtration rate eGFR) and control of blood pressure. The lower the kidney function, the more frequent it should be measured, and for renal function below 60 mL/min/1.73 m2, it has been proposed that measurements are performed with monthly intervals corresponding to kidney function/10 (e.g. every 4 months when creatinine clearance is 40 mL/min).[125] In diabetes with a fast decline in kidney function even more frequent measurements may be needed. Monitoring of renal function and electrolytes (e.g. hyperkalaemia), is clinically important in patients with AF and DM receiving anticoagulation therapy, because declining kidney function increases the risk for bleeding, and the dose of anticoagulants with renal clearance, i.e., DOACs, should be adjusted according to the level of kidney function.[41,125] In the trials comparing DOACs with VKAs, kidney function was evaluated as creatine clearance estimated from creatinine, sex, age, and weight using the Cockcroft and Gault equation (mL/min) (not standardized for body surface area).[152] Therefore, it is recommended to use creatinine clearance (CrCl) when adapting dose of anticoagulation therapy with DOACs.

Benefits of Medical Therapies put Into Perspective

Sub-analyses with respect to patients with and without DM have been performed of all the four Phase 3 DOAC trials.[153] A meta-analysis of these trials demonstrated that DOACs reduce the risk of stroke/SE in patients with DM to a similar extent as in patients without DM. There was no significant modification of the effect of DM on the relative reduction of major bleeding with DOACs vs. warfarin. CV mortality and intracranial haemorrhage were significantly and to a similar extent reduced by DOACS in both the presence or absence of DM.[97] There are however important differences in patient characteristics between the trials.

The proportion of patients with DM in the four Phase 3 DOAC trials studies were between 23.3%, and 39.9%.[84,86,89,154] Patients with DM were on average 0.8–3 years younger than patients without DM, and had a higher CrCl. The CHA2DS2-VASc scores of the included patients importantly differ between trials. In RELY and ARISTOTLE, these differed more than one point between patients with and without DM. Given the fact that DM is one of the determinants of the CHA2DS2-VASc score, the overall risk of stroke with the exclusion of DM was slightly higher in the patients with vs. without DM in RELY and ARISTOTLE.[85,88] Conversely, in ENGAGE, the difference in CHA2DS2-VASc or CHADS2 score was less than one point in DM patients vs. patients without DM.[87,90] Hence, when excluding DM as a risk factor, the stroke risk of patients without DM in RELY and ARISTOTLE was slightly lower than that of DM patients, possibly caused by the higher age. Conversely, when DM is not taken into consideration, there were fewer residual stroke risk factors in patients with vs. without DM in ROCKET and ENGAGE (more than 0.5 points difference in CHADS2 and CHA2DS2-VASc score).

Patients with DM in these four trials had a higher BMI than subjects without DM. Obesity is an important risk factor for CV complications, but in the four DOAC trials, obesity was shown to protect against stroke/systemic embolism. This effect was not demonstrated in observational studies in that same meta-analysis.[155] Others suggest that obesity may increase the risk of thromboembolism and bleeding.[156]

The patient characteristics in these randomized trials contrast with the data from real-world studies. The proportion of DM patients receiving oral anticoagulants increased between 2001 and 2015, and the prescription of warfarin decreased in favour of DOACs.[157] Diabetes mellitus patients with silent AF episodes had a higher risk of stroke irrespective of glycaemic control.[158] In EORP-AF (20.6% of enrolled subjects had DM), DM patients were significantly older than those without DM, had more CKD and CV co-morbidities, and were more often prescribed oral anticoagulants. DM patients had a higher rate of all-cause mortality, CV mortality, and non-CV mortality.[62] A recent analysis of the FANTASIIA (Atrial fibrillation: influence of the level and type of anticoagulation on the incidence of ischaemic and haemorrhagic stroke) cohort in Spain demonstrated that time in therapeutic range in acenocoumarol using subjects, was lower in DM patients compared to patients without DM. In this comparison, DM patients had the same age as non-DM patients, but a much higher CHA2DS2-VASc score, but a higher risk of major bleeding, myocardial infarction, CV mortality, total mortality, but no differences in stroke rates.[83]

In contrast, in ORBIT, where 29.5% of patients had DM and were younger, had a higher BMI more likely to have hypertension, CKD, HF, CAD, and stroke.[31] Slower decline of renal function has been attributed to DOACs compared to VKAs, but this effect was partially lost in patients with DM.[159] Diabetes mellitus was associated with earlier CV mortality in GARFIELD (Global Anticoagulant Registry in the FIELD—Atrial Fibrillation), but the contribution of ischaemic stroke to mortality was low.[160] The presence of DM predicted prescription VKAs over DOACs.[161] In propensity score matched cohorts, DM patients using DOACs had fewer ischaemic strokes and MACE than those using warfarin.[102,162] Conversely, in a tapered matched study, DOACs were associated with more bleeding hospitalizations.[163]

Taken together, there is a large heterogeneity within both the randomized and real-world studies on DOACs for stroke prevention in patients with AF and DM. However, the beneficial efficacy and safety of DOACs compared to warfarin seem conserved in AF patients with DM, irrespective of their baseline stroke risk or the presence of other cardiovascular risk factors.

There is controversy on whether better glycaemic control results affect stroke rates in DM patients. Longer duration of DM has been associated with an increased risk of embolic stroke (but not of major bleeding when treated with oral anticoagulants).[54] Further, in registry studies there is an association between the extent glycaemic control and stroke risk, but vascular damage does not seem to affect embolic risk.[57,63] Diabetes mellitus is an independent predictor for the development of AF.[10,66,164] Glucose lowering therapies decrease the incidence of HF and CKD in DM patients, but in many CV outcome studies AF was not a pre-specified outcome.[165,166] Diabetes mellitus patients with higher Hb1AC values had more ischaemic strokes, but this risk was nullified by the use of oral anticoagulation.[58]