Relationship of Hemoglobin A1C and Mortality in Heart Failure Patients With Diabetes
The cohort was 94% men with a mean age of 69 years and a mean body mass index of 31.7 kg/m2. Mean A1c was 7.75%, and the levels of A1c in each quintile were ? 6.4%, 6.4%-7.1%, 7.1%-7.8%, 7.8%-9.0%, and > 9%. As the A1c quintiles increased, there was a greater prevalence of diabetic complications; individuals were more likely to require a greater number of medications for glycemic control; and they were more likely to be receiving insulin therapy.
Compared with quintile 3, the other 4 quintiles had a significantly higher risk for death over 2 years, ranging from a risk-adjusted hazard ratio of 1.37 (95% confidence interval, 1.14-1.64) for quintile 1, to 1.31 (1.09-1.58) for quintiles 2 and 4, to 1.45 (1.20-1.76) for quintile 5. The risk of heart failure hospitalization increased with larger quintiles, but was not statistically significant after adjustment for potential confounders.
Oops – this is clearly a U shaped of V shaped curve. Lower is not necessarily better.
This finding reminds me a bit of the EPO studies – Epo helped patients, but not enough to achieve a normal Hgb.
In both situations, we must balance the benefits of achieving normal laboratory values with the costs (not financial but medical) of taking more medications. I suspect that the real problem here is that once you achieve a reasonable HgbA1c, the side effects of taking more drugs to achieve a “great” HgbA1c outweigh the benefits of the improved glucose control
Of course diabetologists will criticize this study, or any study that does not fit their medical religion. They will fall prey to my new favorite quote – actually my own quote – “Many diagnostic errors occur because we try to fit the data to our hypothesis rather than fitting the hypothesis to our data.” I submit that the problem here is the same. The hypothesis is that lowering HgbA1c is good – therefore, we must find fault with any study that does not support our hypothesis.
I teach students and residents that we should work hard to get HgbA1c below 8 and that a goal of 7 is a performance target that does not fit the data. I acknowledge the guideline and performance measure, and then look at the data and further state the the Emperor needs some clothes. The current study suggests that we send the Emperor some cloth.
{ 1 comment… read it below or add one }
Sadly, UAB doesn’t have full-text access to this journal. I will have to see if I can read it from work. Excuse me for throwing out anecdotal stuff, but do you find that elderly patients with A1C’s below 7 complain of feeling like crap, compared to patients with less tight control?