• The authors of this registry-based observational study of a Swedish population of patients with type 1 diabetes (N = 33,915) determined the excess risk for death according to level of glycemic control. After a mean follow-up of 8 years, 8% of patients with diabetes had died compared with 2.9% of matched controls. Mortality from cardiovascular causes was 2.7% and 0.9% for each group, respectively.

• The risk for death from any cause and from cardiovascular causes is twice as high in patients with an HbA1c level ≤6.9% and type 1 diabetes compared with the general population.

– Moshe Ornstein, MD

Written by Trevor J Orchard MD
This recent paper by Lind and colleagues is a major contribution to our understanding of the role of glycemic control and cardiovascular disease in type 1 diabetes. It has the strengths of being both contemporary and extremely large, with over 33,000 type 1 diabetic patients (mean age, 35 years) and controls, matched 5:1, from the same national registries. The major finding is that, compared with controls, overall mortality and cardiovascular mortality rates are still two- to threefold higher, even in those with relatively good control (HbA1c <7%), although for both outcomes mortality risk increases in a curvilinear fashion such that, for those with an average HbA1c >9.7%, there is an approximate 10-fold risk for each outcome.

These findings clearly support the concept that high HbA1c is bad in terms of overall and cardiovascular mortality but raise the issue as to whether there is much to gain for reducing the HbA1c below current goals of <7%. For example, the hazard ratios comparing those <7.0% with those <7.9% were 2.36 and 2.38. Furthermore, in contrast with two recent studies from Finland and the US, it seems that mortality in this Swedish study is still increased (twofold), even in those with normoalbuminuria. However, it must be noted that the researchers in the Swedish study followed the patients for an average of 8 years while they had, on average, over 20 years’ duration at baseline before follow-up started. It is therefore highly likely that many of those now classified as good control (HbA1c <7%) may have had significant periods in the past with much poorer control (and possibly had micro- or worse albuminuria, which is now treated to normal levels with ACE inhibitor or ARB medications, which were not included in the analysis).

What clinical lessons are we to conclude from this study? First, it is reasonable to conclude that poor glycemic control plays a major role, especially at high HbA1c levels, but that the relative impact per unit HbA1c gets smaller as one approaches the normal range. The same is true for blood pressure and LDL-C. Secondly, as we get patients in better glycemic control, we should be increasingly looking for other risk factors such as hypertension, insulin resistance, dyslipidemia, and even genetic factors, as these have also been shown to contribute to risk. Thirdly, these results do not— for the reasons stated above—negate the potential value of striving, where appropriate, for good glycemic control (ie, HbA1c <7.0%) and normoalbuminuria. As valuable as observational studies like this one are, the best answer as to the value of intensive therapy for type 1 diabetes, and thereby better control, can only come from a randomized clinical trial. The long-term mortality follow-up of the Diabetes Control and Complications Trial are thus eagerly awaited.

Journal Abstract