Diabetes mellitus (DM) modify outcome in patients with heart failure (HF). We aimed to analyze the risk for death, HF alone, combined end point HF/death, and ventricular tachycardia/ventricular fibrillation (VT/VF) in patients with mild HF without DM and in those with DM, further stratified by the presence of insulin treatment. We determined whether cardiac resynchronization therapy with defibrillator (CRT-D) versus implantable cardioverter defibrillator improves clinical outcomes in these 3 subgroups. Cox proportional hazards regression models were used to analyze 1,278 patients with left bundle branch block in the Multicenter Automatic Defibrillator Implantation Trial With Cardiac Resynchronization Therapy trial. Treatment with CRT-D versus implantable cardioverter defibrillator was associated with 76% risk reduction in all-cause mortality (hazard ratio 0.24; 95% confidence interval 0.08 to 0.74, p = 0.012) in subgroup of diabetic patients treated with insulin only (interaction p = 0.043). Significant risk reduction in HF alone, HF/death, and the VT/VF after CRT-D was observed across investigated groups and similar left ventricular reverse remodeling to CRT-D. In conclusion, patients with mild HF with DM treated with insulin derive significant risk reduction in mortality, in HF, and VT/VF after implantation of CRT-D. Diabetic patients not receiving insulin benefit from CRT-D by reduction of HF events.
Implantable cardioverter defibrillator (ICD) and cardiac resynchronization therapy (CRT) devices have become the standard of care in the treatment of HF. In patients with HF, diabetes mellitus (DM) is an independent risk factor for increased morbidity and mortality, but data are conflicting whether patients with DM derive different benefit from cardiac resynchronization therapy with defibrillator (CRT-D) than those without DM. The effects of insulin on clinical outcome in patients with HF are controversial; observational studies suggest no effect, whereas other studies showed an increased risk for mortality in those receiving insulin. Insulin-treated patients with type 2 DM and a clinical history of cardiovascular disease have increased risk for ventricular arrhythmias. In this study, we aimed to evaluate whether there is an association between insulin treatment and clinical outcome in patients receiving CRT-D in the Multicenter Automatic Defibrillator Implantation Trial With Cardiac Resynchronization Therapy (MADIT-CRT).
Methods
The results and the protocol of the MADIT-CRT trial have been previously reported. From December 22, 2004, to June 24, 2009, a total of 1,820 patients were enrolled at 110 centers in the United States, Canada, and in Europe. Patients of either gender who were at least 21 years old, with ischemic cardiomyopathy (New York Heart Association [NYHA] class I or II) or nonischemic cardiomyopathy (NYHA class II only), sinus rhythm, a left ventricular (LV) ejection fraction (EF) of ≤30%, and prolonged intraventricular conduction (QRS duration ≥130 ms) were randomly assigned in 3:2 ratio to CRT-D or ICD only. The mean follow-up of the enrolled patients was 40 months. Because the benefit of CRT-D in the trial was restricted to patients with left bundle branch block (LBBB), the present study population included 1,278 patients with LBBB (760 with CRT-D and 518 with ICD).
Patients with DM were defined in the original MADIT-CRT protocol as those receiving oral hypoglycemic agents and/or insulin at the time of enrollment. The primary end point of the present study was all-cause mortality. Secondary end points included HF alone, the combined end point of HF or death whichever occurred first, and ventricular tachycardia/ventricular fibrillation (VT/VF) treated by ICD with antitachycardia pacing or shock.
An independent echocardiographic core laboratory analyzed the echocardiography images captured at baseline and 12 months after device implantation. LV volumes were measured by Simpson’s disk method in the apical 4- and 2-chamber views. When analyzing the echocardiographic response to CRT-D, we evaluated changes in LVEF, LV end-diastolic volume, LV end-systolic volume, left atrial volume and LV mass, ventricular septal wall thickness, and ventricular posterior wall thickness. Reverse remodeling effect to CRT-D was defined as percent reduction in the studied parameters between enrollment and 1-year echocardiogram (calculated as the difference between 1 year and baseline, divided by the baseline value).
Baseline clinical characteristics were compared using the nonparametric Kruskal-Wallis test for continuous variables and the chi-square test or Fisher’s exact test for dichotomous variables, as appropriate. We performed Kaplan-Meier survival analyses of unadjusted cumulative event rates stratified by DM and the presence of insulin treatment with the log-rank test for determination of statistical significance. The Cox proportional hazards multivariate regression models were used to estimate hazard ratios for risk of death, HF, the combined end point of HF or death whichever came first, and VT/VF. These hazard ratios were estimated for 3 separate groups: patients without DM, with DM with insulin treatment, and with DM without insulin treatment. The independent variables were chosen using the best subsets selection method. For death, HF, and HF/death, we adjusted CRT treatment for white, presence or absence of ischemic cardiomyopathy, presence or absence of a QRS ≥150 ms, glomerular filtration rate units, and left atrial volume indexed by body surface area (units). Age at enrollment and gender were also forced into the multivariate model to further ensure clinical applicability. The model for VT/VF was adjusted for age at enrollment, female gender, white, glomerular filtration, LV systolic volume at baseline indexes by body surface area (units), systolic blood pressure, NYHA class, QRS >150 ms, and previous ventricular arrhythmias. We followed the statistical methodology because we wanted to develop a parsimonious model, which excluded variables that were not significantly predictive of the end points and would have very little impact on the results. In this way, we attempted to maximize statistical power, an important consideration in subgroup analysis. To assess the CRT-D treatment differences between patients by DM status and insulin treatment, a treatment-by-DM medication interaction term was included in the Cox proportional hazard regression models. A 2 degrees of freedom Wald test was done to assess the strength of the interaction between the 3 DM groups and CRT-D treatment.
To compare the magnitude of changes in the various echocardiographic parameters among the 3 DM groups, the nonparametric Kruskal-Wallis test was used. All statistical tests were 2 sided, and a p <0.05 was considered statistically significant; because of the numerous statistical tests, the p value reported should be considered as nominal and not corrected for multiple comparison. Analyses were carried out with SAS software (version 9.3; SAS Institute, Cary, North Carolina).
Results
In a total of 892 patients without DM (70%), 262 (20%) with DM not on insulin and 124 patients with DM (10%) receiving insulin were included. The mean follow-up was 3.3 ± 0.9 years.
Clinical and demographic characteristics of the study population are presented in Table 1 . Patients with DM treated with insulin did not differ significantly regarding age, body mass index, ischemic etiology of HF, history of hypertension, previous myocardial infarction or coronary artery bypass graft surgery, or regarding the use of cardiovascular medication compared with patients with DM without insulin treatment. However, patients with insulin-treated DM had a lower diastolic blood pressure and glomerular filtration rate compared with those with DM without insulin treatment. Other than smaller LV end-diastolic volume in the insulin-treated DM group, there were no statistical differences in echocardiographic parameters at baseline between groups ( Table 1 ).
| Diabetes Mellitus | |||||
|---|---|---|---|---|---|
| Variables | No (n=892) | Without insulin (n=262) | With insulin (n=124) | p value overall | p-within diabetes |
| Age (years) | 63.5 ± 11.5 | 66.2 ± 9.1 | 65.2 ± 9.4 | <0.01 | NS |
| BMI (kg/m 2 ) | 27.8 ± 4.9 | 29.9 ± 5.5 | 30.5 ± 5.5 | <0.001 | NS |
| Women | 269 (30%) | 80 (31%) | 44 (35%) | NS | NS |
| White | 825 (93%) | 239 (92%) | 101 (82%) | <0.001 | <0.01 |
| CRT-D treatment | 535 (60%) | 165 (63%) | 60 (48%) | <0.05 | <0.01 |
| Ischemic NYHA I and II | 354 (39%) | 139 (53%) | 69 (56%) | <0.001 | NS |
| Non-Ischemic NYHAII | 538 (60%) | 123 (47%) | 55 (44%) | <0.001 | NS |
| Hypertension | 496 (56%) | 206 (79%) | 102 (82%) | <0.001 | NS |
| Prior MI | 249 (28%) | 105 (41%) | 51 (43%) | <0.001 | NS |
| Prior coronary artery bypass graft surgery | 161 (18%) | 81 (31%) | 40 (32%) | <0.001 | NS |
| HF hospitalization | 330 (38%) | 94 (37%) | 62 (51%) | <0.05 | <0.01 |
| QRS (ms) | 163 ± 19 | 161 ± 18 | 161 ± 19 | NS | NS |
| Heart rate (beats/min) | 67.9 ± 11 | 69.3 ± 11 | 68.8 ± 9.2 | <0.01 | NS |
| Systolic blood pressure (mmHg) | 123.0 ± 17.2 | 124.4 ± 17.0 | 122.2 ± 16.3 | NS | NS |
| Diastolic blood pressure (mmHg) | 71.6 ± 10.5 | 71.8 ± 9.5 | 68.8 ± 9.5 | <0.05 | <0.01 |
| Glomerular filtration rate (ml/min) | 70.6 ± 19.9 | 68 ± 20.7 | 63.3 ± 20.9 | <0.001 | <0.05 |
| Baseline drug treatment | |||||
| Angiotensin-converting-enzyme inhibitors/angiotensin receptor blockers | 858 (96%) | 253 (97%) | 117 (94%) | NS | NS |
| Beta blocker | 840 (94%) | 247 (94%) | 114 (92%) | NS | NS |
| Diuretics | 568 (64%) | 201 (77%) | 102 (82%) | <0.001 | NS |
| Baseline echocardiographic parameters | |||||
| LV ejection fraction | 23.4 ± 5.4 | 23.6 ± 5.1 | 23.9 ± 5.2 | NS | NS |
| LV end diastolic volume indexed by BSA | 127.6 ± 31 | 124.2 ± 27.5 | 120.5 ± 29.8 | <0.05 | NS |
| LV end systolic volume indexed by BSA | 91.4 ± 25.3 | 89 ± 21.9 | 86.9 ± 25.1 | NS | NS |
| Left atrium volume indexed by BSA | 47.2 ± 10.3 | 47.3 ± 9.7 | 46.9 ± 10 | NS | NS |
| LV mass | 215 ± 42 | 216 ±37 | 219 ± 37 | NS | NS |
| Ventricle septal wall thickness | 0.82 ± 0.1 | 0.82 ± 0.1 | 0.83 ± 0.1 | NS | NS |
| Ventricle posterior wall thickness | 0.80 ± 0.1 | 0.81 ± 0.1 | 0.82 ± 0.1 | NS | NS |
In multivariate model, in ICD only patients with DM receiving insulin were at significantly higher risk for all-cause mortality (hazard ratio [HR] 2.35; 95% confidence interval [CI] 1.24 to 4.45, p <0.001), for HF (HR 2.16; 95% CI 1.38 to 3.37, p <0.001), and HF/death (HR 2.27; 95% CI 1.5 to 3.43, p <0.001) compared with those without DM. Furthermore, patients with DM receiving insulin were at higher risk of all-cause mortality (HR 2.78; 95% CI 1.19 to 6.45, p = 0.017), for HF alone (HR 1.61; 95% CI 0.96 to 2.70, p = 0.066), and for HF/death (HR 1.70; 95% CI 1.06 to 2.76, p = 0.03) compared with those with DM not treated with insulin. There was no statistical difference for all-cause mortality (HR 1.17; 95% CI 0.54 to 2.51, p = 0.683), HF (HR 1.33; 95% CI 0.88 to 2.02, p = 0.172), and HF/death (HR 1.32; 95% CI 0.89 to 1.96, p = 0.157) in patients with DM not treated with insulin compared with those without DM. Risk for VT/VF was similar among studied groups, in patients with DM treated with insulin versus those without DM (HR 1.27, 95% CI 0.72 to 2.25, p = 0.402), in patients with DM not treated with insulin versus those without DM (HR 1.06, 95% CI 0.66 to 1.68, p = 0.8), and between 2 DM groups (HR 1.2, 95% CI 0.62 to 2.31, p = 0.582).
In patients with LBBB, CRT-D treatment did not reduce all-cause mortality in patients without DM or in those with DM without insulin treatment but significantly improved survival in those with DM receiving insulin (p <0.01; Figure 1 ). CRT-D treatment was associated with a significantly lower cumulative incidence of HF/death (all p <0.001; Figure 2 ), compared with an ICD treatment.
Consistently, multivariate Cox model showed that CRT-D treatment in patients with DM receiving insulin was associated with significant, 76% risk reduction in all-cause mortality, 71% significant risk reduction in HF/death, and 74% reduction for HF alone after adjustment for relevant clinical covariates ( Table 2 ).
| Death | |||||
|---|---|---|---|---|---|
| CRT-D vs. ICD | Patients/Events | HR | CI | p value | P value treatment x diabetes with or without insulin vs. no diabetes |
| No Diabetes | 868/71 (8.1%) | 0.85 | 0.53-1.35 | 0.484 | |
| Diabetes without Insulin s | 257/24 (9.3%) | 0.95 | 0.41-2.17 | 0.9401 | 0.814 |
| Diabetes with Insulin t | 117/19 (16.2%) | 0.24 | 0.08-0.74 | 0.012 | 0.043 |
| Heart failure or Death | |||||
|---|---|---|---|---|---|
| CRT-D vs. ICD | Patients/Events | HR | CI | p value | |
| No Diabetes | 868/180 (20.7%) | 0.54 | 0.40-0.73 | <0.001 | |
| Diabetes without Insulin | 257/70 (27.2%) | 0.42 | 0.26-0.67 | <0.001 | 0.370 |
| Diabetes with Insulin | 117/48 (41%) | 0.29 | 0.16-0.54 | <0.0001 | 0.073 |
| Heart Failure | |||||
|---|---|---|---|---|---|
| CRT-D vs. ICD | Patients/Events | HR | CI | p value | |
| No Diabetes | 868/151 (17.4%) | 0.46 | 0.33-0.63 | <0.0001 | |
| Diabetes without Insulin | 257/56 (21.8%) | 0.32 | 0.18-0.55 | <0.0001 | 0.262 |
| Diabetes with Insulin | 117/40 (34.2%) | 0.26 | 0.13-0.52 | 0.0001 | 0.153 |
| VT/VF | |||||
|---|---|---|---|---|---|
| CRT-D vs. ICD | Patients/Events | HR | CI | p value | |
| No Diabetes | 851/191 (22.4%) | 0.60 | 0.43-0.85 | 0.004 | |
| Diabetes without Insulin | 249/54 (21.7%) | 0.63 | 0.37-1.11 | 0.099 | 0.578 |
| Diabetes with Insulin | 113/20 (26.5%) | 0.26 | 0.10-0.68 | 0.006 | 0.039 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
