As pointed out by Dr. Soo Hoo,8 such single measurements indeed do not capture the dynamics of blood glucose control, and average values can be biased. The sources of bias include different sampling frequencies, and the concomitant presence of hypoglycemic and hyperglycemic events that may result in “normal” calculated averages, whereas the levels of blood glucose are extremely abnormal. The hyperglycemic index, defined as the area under the glucose curve above the upper limit of the target range, divided by the length of ICU stay, also is prone to bias. Ideally, to capture the dynamics of glucose control in an ICU patient, nearly continuous monitoring with an accurate glucose sensor is required, although it remains unclear which aspect of the dynamics (ie, the time spent within the target range, the variance of the blood glucose reading, or the avoidance of hypoglycemia or overcorrection with hyperglycemia following the hypoglycemia) is important to achieve the full benefit of the intervention. We entirely agree with Dr. Soo Hoo that these aspects remain to be elucidated, and the development of continuous glucose sensors is a prerequisite before these questions can be answered.