Cardiac resynchronization therapy with a defibrillator (CRTd) in failing heart patients with type 2 diabetes mellitus and treated by glucagon-like peptide 1 receptor agonists (GLP-1 RA) therapy vs. conventional hypoglycemic drugs: arrhythmic burden, hospitalizations for heart failure, and CRTd responders rate

A base cohort of 579 patients newly treated with non insulin anti-diabetic drugs (metformin, sulfonylureas, thiazolidinediones, glucosidase inhibitors, guar gum, meglitinides, etc.) between 1 January 2010 and 1 January 2017 was assembled. T2DM was diagnosed according to American Diabetes Association criteria [6]. To establish T2DM patients treatment, the screened patients answered a specific questionnaire about medicines used for diabetes treatment, the date of the beginning and end of treatment, route of administration, and duration of use [6]. Therefore, 559 T2DM patients affected by HFrEF were enrolled in the study. The diagnosis of HFrEF was made as indicated by international guidelines on heart failure disease management [2]. After study enrollment phase T2DM patients received a CRT-d treatment according to the international guidelines [2]. Moreover, we randomized the CRTd patients with diabetes under conventional hypoglycemic drug therapy to an GLP1 RA treatment by a computer generating code program. However, 288 CRTd patients with diabetes received GLP-1 RA plus conventional drug therapy, and named as GLP-1 RA group [6]. The remaining patients, 271 CRTd patients with diabetes, did not receive GLP-1 RA therapy, and remained under conventional hypoglycemic drug therapy, and named as conventional group or controls.

Inclusion criteria were: At least 18 years of age, affected by T2DM without insulin anti-diabetic prescription, without GLP1 RA treatment before to receive CRTd, clinical history of stable chronic heart failure, New York Heart Association (NYHA) functional class II or III, left bundle branch block, severe left ventricle ejection fraction reduction (LVEF < 35%), stable sinus rhythm, candidates to receive a CRT-d treatment according to the international guidelines [2].

Exclusion criteria were: Insulin therapy and GLP-1 RA therapy before and at any time of the study, age < 18 or > 75 years, ejection fraction > 35%, previous implant of implantable cardioverter defibrillator (ICD), CRT-d and/or pacemaker, absence of informed patient consent, and any condition that would make survival for 1 year unlikely.

This was an observational multicenter, prospective randomized study conducted at University of Campania Luigi Vanvitelli (Naples, Italy), at Cardarelli Hospital (Naples, Italy), at Catholic University of Sacred Heart (Campobasso, Italy), and at John Paul II Research and Care Foundation (Campobasso, Italy). In this study we enrolled a consecutive population of 560 T2DM patients affected by heart failure (HF) with indication to receive a cardiac resynchronization therapy with defibrillator (CRT-d). After CRT-d, 559 consecutive T2DM patients under conventional hypoglycemic drug therapy were divided in two groups as described before in the text. The clinical characteristics of T2DM patients were well matched and balanced between the two groups of study. Before interventions, the baseline laboratory studies, including HbA1c, lipid panel, and fibrinogen, were determined. Responders patients to a CRT-d treatment were defined by evidence of left ventricle (LV) reverse remodeling, 6 min-walk test improvement and Minnesota living with heart failure scale improvement as previously described [2]. Enrolled patients were followed by clinical, instrumental assessment, and device telemetric control (at implant, 10 days, 6, and 12 months after discharge). During these visits and device interrogations, we reported arrhythmic events, shocks interventions, and subsequently CRT-d effect in terms of clinical outcomes, CRT responder rate, and clinical events as deaths, cardiac deaths, and hospitalizations for HF worsening.

As reported before [5, 6], we defined current exposure to an antidiabetic drug as any prescription whose duration plus a 30-day grace period included the index date (period accounted for non adherence and for the drug’s biologic half-life). For all patients (GLP-1 RA group vs. conventional group), current exposure was classified by the use of the following drugs: GLP-1 RA-based drugs, two or more oral antidiabetic drugs used in combination, a single oral antidiabetic drug. Oral antidiabetic drugs used in combination served as our primary reference category, since GLP-1 RA-based drugs are second-line or third-line therapy and are thus used at a similar point in the management of the disease in failing heart patients [2, 6].

At baseline, and at 6th and 12th month of follow up, a trans-thoracic two-dimensional echocardiograms with M-mode recordings, conventional Doppler, and pulsed-wave tissue Doppler imaging (TDI) measurements was performed in each patient using a Philips iE33 echocardiograph (Eindhoven, The Netherlands). Echocardiographic images were acquired in the parasternal long and short axis views. The LV end-diastolic diameter (LVEDD), end-diastolic volume (LVEDV), end-systolic diameter (LVESD), end-systolic volume (LVESV) were measured, and LVEF was determined with the Simpson method [2]. The amount of mitral regurgitation was calculated as the area of the color-flow Doppler regurgitant jet divided by the area of the left atrium in systole, and described as low (+), moderate (++), moderate-severe (+++), and severe (++++), as previously reported [2, 9]. All echocardiographic studies were performed and analyzed by the same study-independent physicians, blinded to the study protocol. Echocardiographic measurements were systematically averaged in five consecutive samples.

Experienced electrophysiologists in CRT implantation performed the three CRT leads positioning in cardiac chambers, and then connected to the CRT-d generator, as previously described [4]. All CRT-d implant procedures were standardized. Physicians used a multipolar and/or a bipolar LV pacing lead, to reach the target left epicardium vessel, and to have the final LV lead position and pacing configuration, by acceptability of pacing thresholds, absence of diaphragmatic stimulation, and anatomic position (chosen position in the target vessel). The final position of the LV pacing lead was assessed with cine fluoroscopy. Implantation duration was defined as the time between skin incision until suture. We used bipolar LV pacing leads (St Jude Medical, Sylmar, CA, USA; Medtronic, Minneapolis, MN, USA), and quadripolar LV pacing leads (Quartet^® model 1458Q and Promote Q^®, St Jude Medical, Sylmar, CA, USA; Attain Performa^® model, Medtronic, Minneapolis, MN, USA), over-the-wire, steroid eluting with a in-line connector. LV pacing leads were connected to an appropriate bipolar CRT-d device (CRT-d bipolar device, St Jude Medical, Sylmar, CA, USA; Medtronic, Minneapolis, MN, USA), and/or to a quadripolar CRT-d device (Quadra Assura CRT-d device, St Jude Medical, Sylmar, CA, USA; Viva^® QuadXT and Viva^® Quad S cardiac CRT-d, Minneapolis, MN, USA).

In these patients we evaluated physical examination, vital signs, and review of adverse events. For each patient we evaluated body mass index (BMI) as the ratio between weight in kg and the height squared [2]. The CRT-d was monitored during follow up, reporting the functionality of the system, all arrhythmic events, and device interventions.

In all these patients we evaluated, after an overnight fast, the plasma glucose, HbA1c, serum lipids, and B type natriuretic peptide (BNP) by enzymatic assays. In all patients, before intervention, and at follow up, we estimated circulating intact GLP-1, and plasma immunologic active form of GLP-1 [7, 8] by using a specific enzyme-linked immunosorbent assay (ELISA) kit (Active GLP-1 7-36, Epitope). We collected patient’s blood samples in ice-cooled blood collection system for plasma GLP-1 preservation tubes (BD P700), and immediately centrifuged at 2500 rpm for 10 min in refrigerated centrifuge. Samples were stored at − 80 °C. GLP-1 levels (Active GLP-1 7-36, Epitope) measurements were obtained after an overnight fast and after breakfast, as previously reported [6‐8]. We defined as post-prandial GLP-1 values the mean of the four GLP-1 evaluations. In these patients at baseline, and during follow up we measured inflammatory markers.

We evaluated at baseline, and after 12 months follow up, circulating serum levels of pro-inflammatory cytokines (tumor necrosis factor-α, TNF α, interleukin-6, IL6), systemic inflammatory markers (C reactive protein, CRP), and leucocytes and neutrophils count as previously reported [2].

As study endpoints, we monitored in all CRTd patients with diabetes, and in the GLP-1 RA group vs. non GLP-1 RA group of patients the cardiac deaths, all cause of deaths, hospitalization rate for HF worsening, strokes, CRT-d responders rate, arrhythmic burden of sustained events, and ICD shocks. Cardiac deaths, all cause of deaths, hospitalization for HF worsening, and stroke events were evaluated during office follow up visits 10 days after clinical discharge, and after 6th and 12th month by the treating physician, by telephonic interviews, hospital admissions, and discharge schedules [4, 9]. During follow up visits, NYHA classification was re-assessed, and patients graded their overall condition as unchanged or slightly, moderately, or markedly worsened, or improved since randomization by global self-assessment [2, 4, 9]. All patients were instructed regularly to assess body weight, occurrence of dyspnea, and any clinical symptom. At each visit patients were asked whether medical events or symptoms suggestive of cardiac arrhythmias occurred, and an ECG, an ECG Holter monitoring and the device interrogations were both performed to detect the presence of asymptomatic arrhythmias. Clinical evaluations included physical examination, vital signs, and review of adverse events. A fasting blood (at least 12 h from last meal) was performed for biochemical peripheral blood assay evaluation at every visit. CRTd responders patients were identified by clinical and instrumental evaluation as reported by authors [2]. In addition, at each clinical follow-up, arrhythmic burden [atrial fibrillation (AF), ventricular tachycardia (VT), and ventricular fibrillation (VF)] and ICD shocks were evaluated by CRTd interrogations and reported for each patient. AF was defined as an arrhythmia originating from atrial chambers, and classified as paroxysmal, and/or not paroxysmal as previously described [4, 9]. VT was defined as arrhythmia originating from ventricular chambers, classified in sustained and/or not sustained by arrhythmic event duration [9]. VF was defined as a fibrillating arrhythmia originating from ventricular chambers, and associated to hemodynamic instability, and cardiac arrest [9]. ICD shocks were defined as high energy interventions by CRT-d device to restore sinus rhythm during at risk of life sustained VT and or VF events [9].

The study was conducted in accordance with the Declaration of Helsinki. The Ethics Committees of all participating institutions approved the protocol. All patients were informed about the study nature, and gave their written informed, and signed consent to participate in the study. The study was registered in ClinicalTrials.gov, clinical Trial Number NCT03282136.

A qualified statistician analyzed all collected data. The CRT-d and T2DM patients were divided in GLP-1 RA group of patients vs. non GLP-1 RA group of patients (conventional group or controls), and during follow up visits, and controls in CRT-d responders vs. CRT-d non-responders. We postulated that, the number of patients with alterations in primary and secondary endpoints was significantly different between GLP-1 RA group of patients vs. non GLP-1 RA group of patients. Safety analyses were performed on data from all enrolled patients. Continuous variables were expressed as means and standard deviations, and were tested by two-tailed Student t test for paired or unpaired data, as appropriate, or by one-way analysis of variance (ANOVA) for more than two independent groups of data. The categorical variables were compared by Chi square or Fisher exact test where appropriate. Survival analysis was performed with the use of the Kaplan–Meier method. Predictors of the study endpoints were evaluated by using Cox regression models in patients with GLP-1 RA based drugs as compared with oral antidiabetic-drug combinations. A univariate analysis was conducted to examine the association between single principal clinic, echocardiographic, electrocardiographic characteristics, and GLP-1 RA therapy, and 12 months study outcomes. All variables with p value of less than 0.2 in the univariate analysis were subsequently entered into a multivariate model. In the multivariate model, variables were separately selected and a p value of less than 0.05 was considered significant. For all independent predictors, 95% confidence intervals were calculated. Statistical significance was considered for a p value of less than 0.05. The statistical analysis was performed using the SPSS software package for Windows 17.0 (SPSS Inc., Chicago Illinois). We calculated a sample size with 300 participants for each group, with estimated 80% power to detect a change of 0.015 between the mean MPI of the placebo-treated and actively treated groups, at a 5% level of significance. A 20% Loss due to early withdrawals and/or non-evaluable measurements was assumed and, combined with the effect of stratification on analysis, resulted in the requirement to recruit at last 240 patients per treatment group.

In CRTd patients with diabetes the GLP-1 RA therapy in addition to standard hypoglycemic drugs reduced the blood values of BNP at 6th and 12th month of follow up. Conversely, BNP blood values were predictive of hospitalization for HF worsening, and of AF events. BNP is a cardiac peptide relapsed in acute and chronic cardiac stress, and over stretching, and evaluated for the acute and chronic heart failure diagnosis [2, 10]. In fact, BNP measurements provide strong prognostic informations for all cause of deaths, and cardiovascular deaths in patients with heart failure [11]. However, higher BNP blood values are linked to advanced cardiac damage, and cardiac pump failure in patients with chronic HF [10‐12]. Intriguingly, at last one of four patients undergoing CRTd will develop AF events [13]. However, a clear association may exist between AF episodes and BNP levels in CRTd patients. As first, the BNP is released from ventricular and atrial myocytes in response to cardiac wall stress [14]. Indeed, atrial hyper stretching and overload may contribute to elevate the synthesis and relapse of BNP [14], and this is consequently linked to higher rate of future AF events [14, 15]. In addition, in the long term all these adverse mechanical events shifting towards fibrotic processes, may lead to structural alterations of cardiac chambers in CRTd patients [16, 17]. Consequently, this may cause the failure of the cardiac pump in CRTd patients [16, 17], and in CRTd patients with diabetes [18, 19]. An index of the effectiveness of left ventricle pump is the LVEF, that is a parameter calculated by dividing the stroke volume by the end-diastolic volume [20]. However, in the DEFINITE trial the patients with greater improvement of LVEF at follow up experienced a significantly lower incidence of arrhythmic events [21]. On the contrary, a depressed LVEF was associated to worse clinical prognosis, as the result of a higher degree of anatomical ventricular remodeling, and pump failure in failing heart patients [22, 23]. Moreover, the advanced cardiac chambers dilatation and fibrosis may cause a severe depression of LVEF, that associated to the heterogeneity of cardiac repolarization properties, may contribute to an electro-anatomical remodeling [21]. To date, the electro-anatomical remodeling may persist also after the improvement of LVEF induced by CRTd, and it may cause a pro-arrhythmic status and a worse prognosis [21]. In this contest, a higher rate of arrhythmic events and worse prognosis may be seen in patients with longer duration of the QRS interval, as an index of right to left ventricle delay [24, 25]. In fact, the QRS duration is influenced by alterations in ionic channels conduction properties and by ventricular fibrosis, and scar extension, and it is a marker of inter-ventricular dyssynchrony [26]. Therefore, a higher degree of inter-ventricular dyssynchrony is linked to more pronounced electrical-anatomical alterations, and subsequently to higher arrhythmic burden, and worse prognosis in CRTd patients and in CRTd patients with diabetes [25‐34]. In addition, we have to mention that, obese and non-obese patients with mild heart failure have a similar risk also of ventricular tachyarrhythmias, but that this clinical setting does not diminish the benefit of cardiac resynchronization therapy in these patients [35]. Nowadays, in a national cohort of patients eligible for CRTd, nearly 90% received a CRTd device, but unfortunately its use differed by race, implanting operator characteristics and hospital [36]. Moreover, in our study the GLP-1 RA group of patients experienced a reduction of the VT burden about a percentage of 4%, and of the AF burden about a percentage of 40%. In detail, we may speculate that, the T2DM may condition the cardiac ionic channels currents, affecting the depolarizing and repolarizing cardiac activity, and action potentials genesis, duration, and propagation in cardiac walls and heart chambers [9, 37]. In addition, in HFrEF patients with diabetes treated by CRTd, the concomitant anatomic remodeling related to cardiac fibrosis and scar disomogeneity and extension, may also contribute to the increase of the arrhythmic burden [4, 19, 38]. However, GLP-1 RA therapy in addition to standard hypoglycemic drugs may modulate these electrical properties [39], and this may lead to the reduction of the arrhythmic burden in CRTd patients with diabetes [4, 40].

GLP-1 RA therapy in addition to standard hypoglycemic drugs may reduce the endpoint of hospital admission for HF worsening in a percentage superior of 80%, and increasing with a 3.7 fold the CRTd responders rate in HF patients with diabetes. Numerous observations may explain these important study results. From literature data, the acute and chronic GLP-1 RA therapy in addition to standard HF therapy in NYHA class III/IV patients results in a significant improvement of LVEF, of peak oxygen consumption, of 6MWT, and of the quality of life score [39, 40]. To date, in our study we reported for first time in literature, that GLP-1 RA therapy in addition to standard hypoglycemic drugs vs. standard hypoglycemic drugs may reduce hospital admissions, without affecting mortality rate in CRTd failing heart patients with diabetes. Furthermore, patients in the GLP-1 RA group vs. controls experienced an increase in the CRTd responders rate. In fact, the Kaplan curve shows this effect at 12 months of up in GLP-1 RA group vs. controls, by a significant opening of the scissor at follow up end (p value < 0.05) (Fig. 4). The entity of CRTd response is based on clinical measures, as patients symptoms, and functional NYHA class assessment, and by echocardiographic diagnosis of LV reverse remodeling [2]. However, GLP-1 RA therapy in addition to standard hypoglycemic drugs, without affecting mortality may control the glycemic homeostasis in CRTd failing heart patients with diabetes, and by other pleiotropic effects may influence ionic channels properties and functions. This may result in the improvement of the electrical stability of cardiac membrane, and this may favor the genesis and propagation of cardiac potentials though the cardiac cells. All these effects may improve cardiac electrical, and hemodynamic functions also in patients with severe pump failure due to irreversible cardiac fibrosis. At clinical level, this may be translated in the amelioration of cardiovascular hemodynamic, associated to the reduction of the arrhythmic burden, HF symptoms, and NYHA class. Therefore, GLP-1 RA therapy in addition to standard hypoglycemic drugs may increase and enhance CRTd response, as a valuable and unique therapeutic effect vs. the conventional hypoglycemic drug treatments.