We read with interest and furthermore welcome the ideas put forth in the letter by Drs. Knops and Doehner regarding our report “Relation Between Hemoglobin A _1c and Outcomes in Heart Failure Patients With and Without Diabetes Mellitus.” Drs. Knops and Doehner propose that insulin sensitivity, rather than glycosylated hemoglobin (HbA _1c ), may be a better prognostic marker in patients with heart failure (HF) with diabetes mellitus (DM), in light of the inconsistent relation between HbA _1c and outcomes observed in patients with HF and DM in multiple studies. They also suggest that insulin resistance, rather than glucose levels, should be a target of therapy HF patients with DM.

We recognize and have previously reviewed the important role of insulin resistance to the progression of the HF syndrome. Furthermore, we have demonstrated in patients with advanced HF with DM that the insulin-sensitizing agent metformin, previously contraindicated in patients with HF, is in fact associated with improved HF survival. We are currently conducting prospective studies of metformin in patients with HF with DM and pre-DM ( ClinicalTrials.gov identifiers NCT01573949 and NCT01122316 ).

In non-HF populations with DM, hyperglycemia, as indexed by higher HbA _1c , is associated with cardiovascular events and incident HF. Thus, there is an incompletely understood incongruity between HbA _1c in the general diabetic population (an incremental marker of cardiovascular risk) and HbA _1c in the diabetic HF population (an unreliable risk marker). Doehner et al demonstrated that insulin sensitivity as assessed by intravenous glucose tolerance testing using the minimal model technique carries prognostic value in HF; however, this is a rigorous diagnostic test not well suited for clinical use. Whether other indexes of insulin sensitivity, including homeostasis model assessment, may be useful markers or targets in HF requires further investigation.