Utilising Self-Monitored Blood Glucose Data to Further Characterise Glycaemic Control in the ACCORD Trial

Utilising Self-Monitored Blood Glucose Data to Further Characterise Glycaemic Control in the ACCORD Trial
Richard Bergenstal
ADA 2011.

In the ACCORD study there was an increase in mortality related to intensive therapy to control blood glucose. It has been shown that rapid drop in glucose or weight change were not linked to death. There has been evidence published whjich shows that those who experienced 1 or more episodes of severe hypo had increased mortality. What is not clear is the impact of minor hypoglycaemia.
The accumulated SMBG data were analysed to look for minor episodes of hypo. In the intensive group there were 762 versus 683 days (standard treatment group) of SMBG data. In the intensive arm the distribution curve was clearly to the left of the standard treatment group, which is due to a lower glucose throughout the entire 24-hour profiles in the intensive group.
What is also emerging is that when the 24-hour profiles in those who died in either arm are examined there appears to be a greater tendency for the glucose value to drop significantly during the day curve. Analyzed by the number of low blood sugars there were no more lows in those who died versus those who lived. In fact there were more highs in those who died. In the standard arm in those who died there were both more highs and more lows.
Of the standard group who had lots of lows there was a substantial increase in mortality. This suggests that there may be a correlation between swings ion glucose level (from high to low) and mortality. This phenomenon is also seen in the intensive group as those who were swing dramatically above the target glucose were more at risk of dying – ie the more you diverge from the glucose goal you have set, the greater the risk of dying.
In conclusion, severe hypo is a significant risk and mild to moderate hypoglycaemia is probably a risk if the glucose levels are swinging or if the target glucose is in the range set for the standard group (because hypos impleies significant glucose swings).