According to the 2017 ACC/AHA/HFSA, ARNIs (sacubitril/valsartan) were recommended for patients with chronic HF to reduce morbidity and mortality in stable condition, rather than ACEIs or ARBs3. In this study, we investigated the additional effects of ARNI on glycemic control, compared to those of ACEIs or ARBs. We compared HbA1 C level reduction, therapeutic difference to reach HbA1 C target and insulin start time. Over the study period, ARNI did not significantly lower HbA1 C compared to ACE inhibitors or ARBs after adjusting for confounding factors (age, stroke, use of insulin, use of calcium channel blockers, estimated glomerular filtration rate). However, a simple comparison ARNI group showed a significant decrease in HbA1 C at 6, 12 and 24 months compared with ACEIs or the ARB group (p= 0.003, 0.009 and 0.026, respectively). Insulin initiation was delayed in the ARNI group, but this difference was not significant based on the HR result, but the cumulative incidence of insulin initiation was significantly lower in the ARNI group. To our knowledge, this is the first retrospective cohort study to reflect a real clinical setting and to compare the effects of ARNI to those of ACE inhibitors or ARBs on glycemic control.
There was a statistically significant difference in HbA1 C reduction at 12 months between the ARNI group and the ACEI or ARB group regardless of the level of EF (Figure S1 and S2). This is a clinically significant result, even though it did not adjust for confounders. The hypoglycemic effects of neprilysin inhibition occur via modulation of the degradation of several peptides with glucoregulatory properties such as GLP-1, bradykinin, atrial natriuretic peptide (ANP), and B-type natriuretic peptide ( BNP)9. If the activity of neprilysin is inhibited, the plasma concentration of these peptides is increased, which results in a hypoglycemic effect. Another pharmacological mechanism of neprilysin inhibition is increased glucose-stimulated insulin secretion (GSIS)15. It is difficult for ARNI to show dramatic glycemic control. However, ARNI may provide additional hypoglycemic effects in patients with heart failure, whose blood sugar is at the upper limit despite high adherence to antidiabetic medications. Blood sugar control in patients with heart failure is important because diabetes is independently associated with an increased risk of death and rehospitalization8.
However, there was a fluctuation in HbA1 C level, decreasing for 6 months, then rising to its highest value at 18 months, then falling for 24 months (Fig. 2). This could reflect a decrease in overall compliance after 6 months of treatment. In the long-term treatment of diabetes, it is necessary to establish treatment compliance to increase the chances of treatment success. It has been reported that at least 45% of patients with type 2 diabetes do not achieve adequate glycemic control (HbA1 C 16. One of the major contributing factors to this is poor medication adherence.17. In this study, patients with medication compliance greater than 80% were included. However, it was difficult to determine the actual medication compliance because this criterion was calculated according to the number of days of prescription. In addition, patients with heart failure have many concurrent diseases; therefore, there may be problems with polypharmacy, which in turn may lead to poor medication adherence.
In this study, many patients were taking a variety of medications, including diabetes medications (thiazolidinedione, SGLT2 inhibitors, and GLP-1 agonist), which may have influenced changes in blood sugar levels. However, in the stepwise linear regression analysis, diabetes medications were not selected as confounding variables. Confounding variables were stroke, insulin, calcium channel blocker (CCB), age, and eGFR.
Because the EF of patients using ACEI or ARB tended to be higher than those on ARNI, we considered a subgroup analysis based on EF of less than 40%. Similar to the results of all patients, there were significant differences in HbA1 C reduction between two groups at 12 months (p= 0.006, Fig. S2). Therefore, this means that EF was not a factor that enhanced the hypoglycemic effect of ARNI.
In previous studies, ARB and ACEI have already been reported to reduce blood glucose level18.19. When comparing the reduction in HbA1 C among the three ARNI drugs, there was no statistical difference between ARB and ACEI. However, the ARNI group was significantly better than either the ACEI group or the ARB group at all time points except 18 months, suggesting that at least ARNI has the definitive effect on lowering HbA1 C.
In the cumulative incidence of insulin initiation, there was a significant difference between the two groups. However, the ARNI group started insulin at only about one-third that of the ACEI or ARB group with no statistical significance although the hazard ratio was only 0.52.
Although we found a possible glycemic control effect of ARNI, there are some limitations to this study. ARNI is a new drug approved by the US FDA in 2015; thus, the clinical experience is relatively short. Therefore, the relatively small number of patients might have led to statistical insignificance. In addition, incomplete data, such as the lack of periodic monitoring regarding HbA1 Clevel, may also have contributed to the limitations. Since this study is a single-center retrospective study, a large-scale retrospective study of multiple institutions in the future should be conducted by matching these baseline characteristics through a propensity score. However, our study may better reflect the actual clinical setting compared to the post hoc study performed with an RCT12.
Nevertheless, this study presents the possibility of a hypoglycemic effect and a delay in the progression of DM when ARNI is used in heart failure patients with type 2 DM. In the future, further studies are needed. to define the hypoglycemic effect of ARNI on a large number of patients.