Controversies about tight glucose control

Following the study of Van den Berghe et al. that was published in 2001, the possibility of hyperglycemia being a mediator, rather than a marker, of critical illness emerged [1]. Simultaneously, new insights regarding the cellular mechanisms of glucose toxicity have demonstrated a possible link between glucose, cytopathic hypoxia and the production of reactive oxygen and nitrogen species [2,3]. However, as reviewed extensively by Corstjens et al., the ultimate proof that hyperglycemia is an independent risk factor for poor outcome in critically ill patients is not currently available [4]. Nonetheless, since the publication of the first Leuven study article in 2001, several different teams have tried to reproduce the results and investigated the underlying mechanisms of the impressive findings of the Leuven team. From the work of these numerous researchers, several controversies and important questions for physicians taking care of critically ill patients actually emerged; most of these are still unanswered. Some of these controversies (i.e., the optimal value of blood glucose, the risks associated with hypoglycemia, the variability of blood glucose and the definition of the categories of patient who could benefit from tight glucose control by intensive insulin therapy) will be discussed briefly in this editorial.

What is the optimal value of blood glucose during critical illness? In experimental conditions, concentrations of glucose higher than 15 mmol/l were clearly deleterious. As these concentrations are irrelevant to the situation of critically ill patients, only clinical data can be used to define the optimal value for tight glucose control. Physiologically, ‘stress hyperglycemia’ is known as an essential component of the host’s response, as a result of the resistance to insulin designed to provide glucose to injured tissues. Therefore, the ‘normoglycemia’ of critical illness is unknown [5]. However, based on the data of the two Leuven studies [1,6], a blood glucose higher than 10 mmol/l can probably no longer be considered as an acceptable target for insulin therapy in critically ill patients. Consistently, three retrospective trials found that patients in whom blood glucose was below 8 mmol/l had a better outcome compared with higher concentrations [7–9]

However, large prospective trials of tight glucose control by intensive insulin therapy comparing two ranges of blood glucose are needed in order to define the optimal target. The Normoglycaemia in Intensive Care Evaluation and Survival Using Glucose Algorithm Regulation (NICE-Sugar) and the Glucontrol trials were designed to compare the effects of insulin therapy titrated to restore and maintain blood glucose between 4.4–6.1 versus 7.8–10.0 mmol/l while the Efficacy of Volume Substitution and Insulin Therapy in Severe Sepsis (VISEP) trial compared 4.4–6.1 to 10.0–11.1 mmol/l in septic patients. The Glucontrol trial was stopped recently after the inclusion of more than 1100 patients. In spite of a satisfactory achievement of the target range, there was no significant difference in the vital outcome between the two groups. The rate of hypoglycemia and the mortality in the patients having experienced at least one episode of hypoglycemia (defined as a blood glucose below 2.2 mmol/l) were higher compared with in patients who did not experience hypoglycemia [10]. Tight glucose control is associated with a four- to sixfold increase in the incidence of hypoglycemia, which is a critical issue, as it represents the major fear and extensive workload when starting intensive insulin therapy [11]. By contrast, in both Leuven studies [1,6], the condition of the patients who experienced hypoglycemia did not worsen, compared with individuals who did not present any hypoglycemic episode. This does not exclude the possibility that long-lasting hypoglycemia may be deleterious or even life threatening as a consequence of decreased glucose availability for vital tissues unable to use other sources of energy, such as in case of ischemia as injury (myocardium, brain). Clearly, an accurate understanding of the consequences of hypoglycemia in critically ill patients requires further investigation.

Meanwhile, the available data from the VISEP and Glucontrol trials [10,12] suggest that the risk-to-benefit ratio of a target blood glucose between 7.8 and 10 mmol/l is more advantageous than the lower target assessed. Even though a further confirmation of these findings is desirable, most intensive-care physicians presently use this intermediate range as a target for intensive insulin therapy [13–15].

Another important question was raised by Egi and coworkers [16]. These investigators observed that, in a retrospective data collection in a set of 7049 critically ill patients, the coefficient of variability calculated from the standard deviation of the values of blood glucose recorded in each patient appeared to be closely related with survival, which is consistent with similar findings in diabetic patients. Outside the intensive-care unit (ICU), recent data recorded in diabetic patients and compared with volunteers also indicated that glucose fluctuations increased the oxidative stress, measured by the urinary concentration of 8-iso PGF_2α, an index of the peroxidation of eicosanoids [17]. These clinical data can reflect the ‘cellular’ data where cell damage was the most prominent when the glucose in the culture medium changed rapidly from a normal to an elevated level (reviewed in [18]). This potentially important issue of glucose variability was not analyzed in the large trials performed in critically ill patients and already published [1,5–8]. In the Glucontrol study, the standard deviation of blood glucose was identical in the two treatment arms. In another recent study [19], a correlation between the glycemia lability index and mortality in patients with sepsis was reported. The current implication that could follow the discovery of the importance of keeping glucose as stable as possible is to boost the use of strict algorithms to maintain glucose within a narrow range. Comparisons between algorithms should then include the comparison of indices of variability [20,21].

Another matter of debate is the choice of patients that would benefit from tight glucose control. In the medical ICU Leuven study [6], the improvement in vital outcome was only seen in the patients who stayed in the ICU for a period of longer than 3 days, while the mortality tended to increase in the short stayers randomized to the intensive insulin therapy group. Another emerging category of patients that would not benefit from intensive insulin therapy is the subset of patients with pre-existing diabetes, as shown by the aggregation of the results of both Leuven studies [22]. The maintenance of a blood glucose level that remains as stable as possible in diabetic patients seems to represent a critical factor [16,17].

The other retrospective large studies [7,9] were performed in mixed populations and concluded that the maintenance of blood glucose below a threshold of 8 mmol/l was associated with an improved outcome, regardless of the category of patients.

Therefore, at the present stage, there is no definite answer to the question of which subgroup is prone to benefit more from intensive insulin therapy than other subgroups. Patients presenting with myocardial ischemia and after cardiac surgery may represent a subset of patients particularly susceptible to the deleterious effects of hyperglycemia ([23] for review). Interventional studies performed in specific subgroups are, therefore, definitely needed.

In conclusion, several important questions are pending before a formal recommendation can be suggested. Titration of insulin in order to prevent an increase of blood glucose over 10 mmol/l and frequent blood glucose checks to minimize the variability of glycemia and the risk and duration of hypoglycemia are currently supported by the available clinical data.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.