Cardiac adiposity and arrhythmogenesis in patients with type-2 diabetes mellitus

Impact of type-2 diabetes mellitus in the outcomes of catheter ablation of atrial fibrillation Rationale Catheter ablation is a well-established treatment for AF, however data regarding efficacy and safety in the DM population are mainly restricted to small size and single centre reports, with conflicting results [125] and limited use of the cryoballoon technique. The aim of this study was to further investigate implication of DM on the outcomes of catheter ablation of AF. Methods This was an observational non-randomised study conducted in 7 European centres. We included all patients aged over 18 undergoing a left atrial ablation procedure during a 24 months' time interval, with AF refractory to at least one class I or class III antiarrhythmic drug. All patients provided written informed consent prior to the procedure. We assessed DM as a potential independent predictor of AF/atrial tachycardia relapse. The study complied with the Declaration of Helsinki and the research protocol was approved by the local ethics committees. Demographics, and admission day anthropometric data were collected. Patients' notes and electronic records were systematically assessed to identify relevant comorbidities. Data from the referral transthoracic echocardiogram was analysed and a multislice computed tomography scan imaging of the left atrium was systematically collected pre-procedure. Procedures were performed under sedation or general anaesthesia, according to each institution's protocol. Venous access was obtained via the femoral vein, with use of vascular ultrasound at operator's discretion. In the absence of patent foramen ovale, a single or dual transseptal puncture was performed under fluoroscopic guidance. Use of transoesophageal echocardiography was left at operator preference. Intravenous heparin was administered to maintain an activated clotting time of 300–350 seconds. Pulmonary vein isolation was the main procedural endpoint and was performed as a first step in all procedures. If the patient was in AF at the start of the procedure and the arrhythmia organized into an atrial tachycardia this was mapped and ablated. In patients undergoing cryoballoon ablation, if the patient remained in AF after isolation of all four pulmonary veins, direct-current cardioversion to sinus rhythm was performed and no further ablation undertaken. In patients undergoing radiofrequency ablation of persistent AF and not cardioverting to sinus rhythm or not organizing to atrial tachycardia during ablation we mapped and ablated areas of complex fractionated atrial electrograms in both atria and the coronary sinus and subsequently DC cardioverted the patient if AF persisted. Patients were evaluated at 3, 6, and 12 months after the procedure. A 12-lead electrocardiogram (ECG) and 24-hour ECG Holter were performed at each follow-up. Additional patient visits and further testing were allowed in case of symptoms. Annual follow-up was performed after the first year. Antiarrhythmic drugs at discharge were prescribed only in selected patients (i.e., for those suffering from longstanding persistent AF, or having relapse during the admission, needing for cardioversion, etc.) and at the operator's discretion, and were discontinued after the first 3 months in the absence of recurrence. Blanking period was defined as the first 3 months post-procedure. Recurrence was defined as any symptomatic or asymptomatic atrial arrhythmia lasting >30 seconds following the blanking period. In case of arrhythmia recurrence during the blanking period not responding to cardioversion (pharmacologic or electrical), patients were classified as having a relapse. The main efficacy endpoint was freedom from atrial arrhythmias following a blanking period of three months. AF or atrial tachycardia relapse during the initial 3-month blanking period was also documented. The following complications were systematically recorded in order to explore safety: vascular complications (if requiring intervention or prolongation of admission), thromboembolism (transient ischemic attack, stroke and/or systemic embolism during or in the first month after the procedure), phrenic nerve palsy, pericardial effusion (if causing haemodynamic instability and/or requiring pericardiocentesis or prolonged monitoring), oesophageal fistula, and procedure-related death. Other complications were reported at the discretion of the operator. The chi-square test was used for categorical and Student's t-test for continuous variables. Levene's test was used to check the homogeneity of variance; equivalent non-parametric tests were used when Kolmogorov–Smirnov was in favour of the absence of normal distribution. Results with P <0.05 were regarded as significant. Kaplan-Meier curves were traced for illustrating freedom from AF or atrial tachycardia among patients with or without DM, and the log rank P test was used for assessing existing differences. Independent predictors of sinus rhythm maintenance after a single ablation procedure were assessed through Cox regression (Method: Forward Likelihood Ratio, Probability for Stepwise 0.05). A propensity score matching was performed to adjust for differences in baseline clinical characteristics. A propensity score was obtained through binary logistic regression: diagnosis of DM (yes or no) was the binary outcome, and all baseline variables were used as covariates for estimating a probability (the propensity score). Then, probabilities in the DM group were matched 1:2 to the closest non-DM patient fulfilling inclusion criteria using the nearest neighbour matching approach. The propensity score was matched to 5 decimals whenever possible. If this was not possible, we subsequently attempted 4, 3 and then 2 decimal matching. If a DM patient could not be matched to any non-DM patient on the second digit of the propensity score, then the DM subject was discarded from the matched analysis. Comparisons between DM and no-DM were performed. Based on Stuart [126], analyses were performed using the groups as a whole, rather than using the individual matched pairs. PASW Statistics version 18.0 was used for descriptive and inferential statistical analysis. Results A total of 2504 patients (mean age 61.1 ± 10.2, 29.4% female) underwent catheter ablation of AF. As many as 234 patients (9.3%) suffered from DM. Most patients had paroxysmal AF (57.5%) at baseline, and mean AF duration was 5.0 ± 5.4 years. The cryoballoon technique was adopted in 29.4% of the patients. Pulmonary vein isolation was achieved at the end of the procedure in almost all the patients (99.0%), with no significant differences between the two groups. Use of the cryoballoon was comparable among patients with or without DM. Rate of relapse during blanking was significantly more frequent in subjects with DM (24.3% vs. 32.8%, p = 0.012). Similarly, relapses at 12 months occurred more frequently in the DM group (25.3 vs. 32.0%, p = 0.031) (Table 4). After adjusting for type of AF (i.e., paroxysmal vs. persistent), during a median follow-up of 17 ± 16 months, atrial arrhythmia free-survival was lower in the diabetics vs. non-diabetics after ablation of persistent AF (log-rank p = 0.003), and comparable after ablation of paroxysmal AF (log-rank p = 0.554). These findings were confirmed after comparing the DM patients vs. a propensity-matched group of non-diabetics (log-rank p = 0.038 for persistent AF). Assessment of independent predictors of AF or arrhythmia relapse is illustrated in Table 5. On multivariate Cox regression, DM, BMI, AF duration and LA volume were independent predictors of relapse. The rate of peri-procedural complications was similar among DM and non-DM patients (3.8% vs. 6.4%, p = 0.128).The incidence of cardiac tamponade, other bleeds, major vascular complications, phrenic nerve palsy, and stroke, transient ischemic attack or systemic embolism was very low and comparable. Efficacy and safety of cryoballoon ablation were comparable to radiofrequency ablation in both DM and no-DM groups (log-rank p = 0.437 for persistent AF and p = 0.531 for paroxysmal AF). Discussion The main finding of this multicentre study is that DM is associated with a higher incidence of atrial arrhythmia relapses at 12 months in patients undergoing catheter ablation of AF. On the Kaplan-Meier analysis, after adjusting for type of AF (i.e., paroxysmal vs. persistent), arrhythmia-free survival at a median follow-up of 17 ± 16 months was lower in diabetic patients with persistent AF compared to those with no DM; however, relapse rates were similar in DM vs. non-DM subjects undergoing ablation for paroxysmal AF. In our series, patients with DM have a higher prevalence of comorbidities such as hypertension, obstructive sleep apnoea, vascular disease and congestive heart failure, and more commonly suffer from non-paroxysmal forms of AF. However, on a multivariate analysis, after adjusting for confounding factors, DM remains an independent predictor of atrial arrhythmia relapses (HR1.39; CI95%1.07-1.88; p = 0.016). The higher rate of post-ablation relapses in the DM population was also confirmed in a propensity-matched analysis. Finally, catheter ablation of AF appears to be safe in DM patients, with no significant difference in the complication rate compared to the non-diabetics. Notably, despite DM being a risk factor for thromboembolism in the AF population, we have found no differences in the number of peri-procedure thromboembolic events between diabetics and non-diabetics. Efficacy and safety of the cryoballoon ablation was comparable to radiofrequency ablation in both DM and no-DM group. Our findings are clinically relevant, considering the high prevalence of DM and its strong association with AF. Subjects with DM have not only an increased risk of developing AF but are also more prone to AF-related complications such as thromboembolism and heart failure. For these reasons, DM patients might warrant the greatest benefit from an effective treatment of this arrhythmia, with the potential aim not only to improve quality of life, but also prevent its relevant clinical sequalae. The present study confirms that AF ablation is effective and safe in the DM population, despite this traditionally representing a higher risk sub-group with more frequent comorbidities. In fact, among DM patients, as much as 80.2% of those with paroxysmal AF and 57.6% with persistent AF were free from atrial arrhythmia at the 12 months' follow-up. However, an important finding of our analysis is that DM is associated with higher long-term relapse rate after catheter ablation of persistent AF, while outcomes for paroxysmal AF are similar among diabetics and non-diabetics. DM is known to cause significant myocardial remodelling (i.e., diabetic cardiomyopathy) and can promote AF through several physiopathological mechanisms, as discussed above. It is conceivable that in diabetic patients, compared to the non-diabetic, persistent forms of AF are associated with a more severe degree of atrial myopathy and a more complex and multifactorial substrate, resulting in a lower long-term efficacy of catheter ablation. Indeed, DM has been independently associated with left atrial enlargement, regardless of concomitant hypertension and diastolic dysfunction [127]. Our finding could have relevant clinical implications, as an early ablative strategy might be particularly valuable in subjects with DM, in order to prevent the progression from paroxysmal to persistent forms of AF, as the latter appear to be more aggressive and difficult to treat. The impact of DM on the outcomes of AF ablation has been previously evaluated by other authors, with conflicting results [125]. In the absence of randomised trials, to the best of our knowledge, the largest available controlled study included 339 DM patients from the German Ablation Registry [128]; in this series, after a median follow-up of 460 days, no differences were found between subjects with or without DM in terms of arrhythmia-free survival. However, these results included diabetic patients with both paroxysmal and persistent AF, and as such no separate outcomes were provided for subjects with different forms of AF; in addition, patients with persistent AF were underrepresented, and results were based on telephone follow-up only. A systematic review and metanalysis by Anselmino et al [129] showed similar outcomes of AF ablation in DM patients compared to the general population, although with relatively frequent need of redo procedure in the diabetics. However, data from a metanalysis including 886 individuals should be interpreted very carefully, especially in the context of relevant methodological bias such as the absence of a direct comparison with a control group. Another relevant finding is that cryoballoon ablation appears to be effective and safe in DM patients, showing comparable results with the radiofrequency technique. These findings are of interest, as the diabetic population was underrepresented in most of the studies evaluating cryoballoon AF ablation; as such, the cornerstone FIRE and ICE trial included only 22 and 37 diabetic patients in the radiofrequency and cryoballoon group, respectively [130]. Finally, the results of the present study might suggest a potential benefit of an adequate treatment of DM to counteract its deleterious effect on the long-term outcomes of AF catheter ablation. The ARREST-AF study demonstrated that an aggressive risk factor management, including better glycaemic control, significantly improves arrhythmia-free survival after catheter ablation of AF [131]. However, although promising, these data should be confirmed in a prospective randomised fashion. Several limitations should be acknowledged. First, no data regarding glycaemic control (e.g. HbA1c), DM duration and therapy were available. In addition, this was a multicentre study including experienced large volume centres and might not represent the type of ablation activity performed in other centres with lower caseloads. Obstructive sleep apnoea has emerged as an important predisposing factor for AF, however patients in the present study were not systematically screened for this condition; therefore, the prevalence of obstructive sleep apnoea in our population could be underestimated. Finally, systematic use of implantable loop recorders may have allowed to document a higher rate of asymptomatic recurrence. Conclusion Catheter ablation of AF appears to be safe in patients with DM. However, DM is associated with higher rate of atrial arrhythmia relapse after catheter ablation, particularly for patients with persistent AF. Cardiac steatosis and atrial fibrillation in patients with type-2 diabetes mellitus Rationale In recent years, oxidative stress and inflammation have been shown to be central mediators of AF in diabetic and obese patients [132]. Antioxidants agents have been shown to reduce atrial remodelling in animal models [19]. Furthermore, there are evidences of increased level of reactive oxygen species in "mitochondria isolated from atrial tissues of diabetic patients and animal models" [102]. Additional sources of reactive oxygen species in the diabetic heart include "Xanthine oxidase, NADPH oxidase, Monoamine oxidase, Protein Kinase C, Nitric oxide synthase (NOS), and Advanced glycation end-products (AGE)" [102]. Moreover, "antioxidant defence systems such as glutathione are depleted in the atria of diabetic hearts" and "mismatch between reactive oxygen species scavenging and generation promotes oxidative stress and inflammation" [102]. Cardiac steatosis can cause lipotoxicity, which is a central pathway of oxidative stress [3]. There are evidences suggesting a correlation between epicardial fat and AF, however the role of cardiac steatosis (i.e., intramyocardial fat infiltration) in the pathophysiology of AF in DM patients has not been investigated. The aims of this study were: 1) to compare the amount of in vivo cardiac steatosis between DM patients with or without background of AF and 2) to characterize oxidative stress and its correlation with in vivo cardiac steatosis in DM patients with or without background of AF. Methods Patients with type-2 DM and coronary artery disease undergoing elective coronary artery bypass surgery (CABG) were included in the study. The study complied with the Declaration of Helsinki and the research protocol was approved by the Barts BioResource review committee (registration ID number 77). All the patients provided written informed consent. In the AF group, participants had a previous established diagnosis of AF with confirmed rhythm documentation using an electrocardiogram (ECG). ECG, transthoracic echocardiogram, and routine blood test (including glycated haemoglobin and fasting blood sugar) were routinely performed pre-surgery in each patient. Anthropometric variables including height and weight were collected, and body mass index was calculated. Exclusion criteria included age <18, history of congestive heart failure with reduced ejection fraction, cardiomyopathy, hemodynamically significant valvular heart disease, valvular AF, previous cardiac surgery, thyrotoxicosis, and inability to provide informed consent. To avoid confounding factors, women were not included given previous evidence of sex dimorphism in the cardiac adiposity distribution and function [3, 21]. During cardiac surgery, myocardial specimen from the right and/or left atrial appendage were obtained in the operating theatre. The samples were collected and stored at -80 degrees. Primary rabbit polyclonal anti 4-hydroxynonenal and rabbit polyclonal anti-ADFP (perilipin 2- PLP2) were used at dilution of 1:100. The 4-hydroxynonenal is a highly reactive product of lipoperoxidation and was used as a marker of oxidative stress [133]. The PLP2 was used to quantify the amount of cardiac steatosis [21]. The immunohistochemistry was performed using the Ventana DabMap Horseradish Peroxidase Kit. The expression of PLP2 and 4-hydroxinonenal was assessed using a semiquantitative scoring (0: absent, 1: slight, 2: moderate, 3: intense) on 10 randomly selected areas. Based on a previous study which quantified a PLP2 optical density of 0.24 ± 0.05 [21] and assuming a 30% higher density in patients with background of AF, a sample size of 16 patients (8 in both groups) was estimated with alfa 0.05 and 80% power. The chi-square test or Fisher's exact test for categorical variables and Student's t-test for comparison of means. Levene's test was used to check the homogeneity of variance; equivalent non-parametric tests were used when Kolmogorov–Smirnov was in favour of the absence of normal distribution. Pearson's coefficient correlation was implemented for testing association between the variables. Results with P <0.05 were considered significant. SPSS version 26.0 was used for statistical analysis. Results A total of 16 patients (68.9 ± 9.5, 100% men) were enrolled, 8 with history of AF and 8 with no previously diagnosed AF. In the AF group, 7 subjects (87.5%) had persistent AF and one paroxysmal AF. All patients in both groups suffered from hypertension and hypercholesterolemia. The mean BMI was 28.4 ± 13.3 kg/m2, with no significant difference between the two groups. The mean left ventricular systolic function was 47.2 ± 12.4 vs. 52.5 ± 13% in the AF and non-AF group, respectively (p = 0.21). As many as 6 patients in the AF group were on metformin, one was on insulin only, and one on no medical therapy for DM. Among those with no AF, 6 were on metformin, one on metformin and gliclazide, and one on sitagliptin. Detailed baseline population characteristics are shown in Table 2.1. PLIN2 was significantly higher in patients with history of AF compared to those with no AF (1.9 ± 0.3 vs. 1.4 ± 0.5, p = 0.04). 4-hydroxynonenal was slightly higher in patients with AF vs. no AF, however the difference was not statistically significant (2.0 ± 0.5 vs. 1.5 ± 0.5, p = 0.08). A Pearson's regression analysis demonstrated a moderate correlation of PLIN2 expression with BMI (r = .53) and 4-hydroxynonenal expression (r = 0.35), and a low correlation with LA size (r = 0.22). There was a negative correlation of PLIN2 expression with age (r = -0.2) and left ventricular ejection fraction (r = -0.12). Discussion The main finding of the present study is that intramyocardial fat (i.e., cardiac steatosis) appears to be higher in diabetic patients with history of AF compared to a control group with no AF. To the best of our knowledge, this is the first study to investigate the correlation between cardiac steatosis and AF. There is a growing body of evidence showing a role of cardiac adiposity in the pathogenesis of AF, however previous reports have mainly focused on the study of epicardial fat. Cardiac steatosis has more recently emerged as a possible contributor to several myocardial diseases [3, 21-22]. A recent study by Mazzali et al [21] has shown a higher concentration of intramyocardial fat in patients with vs. without coronary artery disease. Furthermore, cardiac steatosis has been associated with deterioration of left ventricular diastolic function and increase of left ventricular mass [21-22]. From a physiopathological perspective, we hypothesize that cardiac steatosis might contribute to the development and perpetuation of AF by promoting oxidative stress. Previous studies have implicated increased oxidative stress within the atrial tissue in the pathogenesis of AF [134-138]. Our data seem to corroborate this hypothesis by showing a trend toward a higher concentration of 4-hydroxynonenal in DM subjects with AF vs. no AF, although the difference was not statistically significant; the small sample size might account for the lack of significant difference. Intramyocardial fat accumulation causes a hyperactivation of the beta-oxidation with subsequent excess formation of reactive oxygen species, which mediate lipotoxicity. Mihm et al first demonstrated an increase of nitrotyrosine and protein carbonyl formation in the right atrial appendage of patients undergoing cardiac surgery with vs. without AF, providing evidence of oxidative damage in human AF [134]. In keeping with these findings, Corradi et al demonstrated increased atrial tissue levels of hemeoxygenase-1 and 3-nitrotyrosine in patients with persistent AF compared with controls [135]. Oxidative stress can promote arrhythmogenesis by several mechanisms [136]. Excessive reactive oxygen species have been shown to modify "several ionic currents in cardiomyocytes, cardiomyocyte coupling, and extracellular matrix" [136-137]. Specifically, oxidative stress prolongs the duration of the action potential and favours triggered activity (early after-depolarisation) by promoting late Na+ current [136-137]. Furthermore, reactive oxygen species reduce total Na+ channels causing a delay in the cardiac conduction and therefore promoting reentry [136-137]. In addition, oxidative stress can stimulate L-type Ca2+ currents with subsequent prolongation of the action potential duration and reduction of the repolarization reserve [136-137]. Reactive oxygen species can induce inflammation and promote cardiac fibrosis and also interfere with connexin proteins in the gap junction prolonging cardiac conduction [136-137]. Moreover, the Ca+2/calmodulin-dependent protein kinase II has emerged as a reactive oxygen species-activated proarrhythmic signal, which can trigger AF via Ca+2 leak through ryanodine receptor [136, 138]. The findings from the present study might have relevant clinical implications. Non-invasive quantification of cardiac steatosis using proton magnetic resonance spectroscopy might be adopted to predict the risk of developing AF in DM patients. AF is associated with worse cardiovascular outcomes and increased risk of death and hospitalization in the DM population [100-101], hence identification of subjects at higher risk of AF could prompt a more aggressive management of DM and other risk factors. In addition, measuring cardiac steatosis in DM patients might potentially guide AF treatment and predict outcomes of catheter ablation. Indeed, our data suggest that cardiac steatosis is an important element of the atrial substrate remodelling in the DM population, and such structural changes might contribute to explain the lower efficacy of AF ablation in diabetics vs. non-diabetics, as demonstrated in our previous study. For these reasons, the presence of cardiac steatosis might potentially help to identify DM subjects with a more advanced atrial substrate who are therefore less likely to benefit from catheter ablation. However, further studies with larger sample size are required to confirm this hypothesis. On the other side, pharmacological treatments targeting intramyocardial fat infiltration and/or oxidative stress might have an important role in the management of AF, particularly in the DM population. Vitamin C has strong antioxidative capacity and has been shown to reduce the risk of postoperative AF in patients undergoing cardiac surgery [139]. However, data appear to be contradictory overall [140] and evidences beyond the setting of postoperative AF are limited. N-3 polyunsaturated fatty acids (omega 3) have been proposed as indirect antioxidant agents; indeed, by inducing a low-to-moderate increase in radical oxygen species levels they could decrease the vulnerability of myocardial tissue to a subsequent oxidative challenge [141]. A combination therapy with omega-3 fatty acid, vitamin C and E has been shown to reduce postoperative AF in a recent metanalysis, while omega 3 alone did not show any significant effect [142]. Statins have been extensively investigated for reduction of AF, mainly in the postoperative setting. Beyond the lipid lowering action, statins are well known to have pleiotropic effects, including antioxidant [143-144]. A recent study has found that statins reduce the amount of epicardial fat [143], but whether they have a similar effect on cardiac steatosis is currently unknown. The role of statins in the prevention of AF remains unclear. Although statins reduce the incidence of postoperative AF in patients undergoing cardiac surgery [145], a recent metanalysis has not identified any impact of statin in the incidence of AF recurrence post catheter ablation [144]. Of note, to the best of our knowledge there are no studies specifically designed to evaluate the impact of statins or antioxidant drugs in the DM population with AF. Given the important role of cardiac adiposity and oxidative stress in DM patients with AF, it is conceivable that this population might particularly benefit from antioxidant and lipid lowering therapy. In the absence of available data, further research is required to address this relevant topic. With regards to the antidiabetic drugs, in a study including 74 participants with type-2 diabetes assigned to metformin or pioglitazone or placebo, no changes in the myocardial triglyceride content was demonstrated after 24 weeks [21]. Sulfonylureas and insulin have been shown to have a pro-AF effect through induction of hypoglycaemia [146]. A possible anti-arrhythmic effect of metformin and GLP-1 receptor agonists has been suggested, but further studies are required to confirm these findings [146-147]. It remains unclear whether thiazolidinedione and DPP-4 inhibitors might have any anti-arrhythmic property [146]. SGLT2 inhibitors appear as the most promising drugs for improvement of cardiovascular outcomes and have been shown to reduce the risk of AF in the DM population, as confirmed by a recent metanalysis [148]; however, the underlying mechanisms have not been entirely understood. A recent study on animal models found that empaglifozin can reduce the late sodium current-induced calcium overload, and also attenuate the level of oxidative stress [149]; however, further research is required to confirm these findings. Conclusion Cardiac steatosis appears to be higher in diabetic patients with history of AF compared to a control group with no AF. Intramyocardial fat infiltration might contribute to the onset and perpetuation of AF in DM patients by promoting oxidative stress. Non-invasive quantification of cardiac steatosis in the DM population might allow to identify subjects at higher risk of developing AF, who might benefit from a more aggressive treatment for preventing cardiovascular sequelae.