Continuing controversy in the intensive care unit: why tight glycemic control, nutrition support, and nutritional pharmacology are each necessary therapeutic considerations

This issue focuses on the latest US [1] and European [2] guidelines, energy delivery, and glycemic control issues. Important transatlantic differences between the two parenteral and enteral nutrition societies remain, which Singer et al. (pp. 170–176) factually describe, while also insisting on the strong rationale behind the European approach. The American Society for Parenteral and Enteral Nutrition (ASPEN)/Society of Critical Care Medicine (SCCM) experts hesitate to recommend the administration of parenteral nutrition to nonmalnourished intensive care unit (ICU) patients receiving some, but not adequate, amount of enteral nutrition during the first 7–10 days after admission. In contrast, European Society for Clinical Nutrition and Metabolism (ESPEN) guidelines focus on preventing an early energy debt by preferentially providing enteral nutrition first but supplementing with parenteral nutrition if target needs cannot be met by ICU day 4. These recommendations are mainly based on observational studies showing a strong correlation between negative energy balance and morbidity/mortality. Ongoing prospective studies should clarify this issue when results are available. Thibault and Pichard (pp. 177–183) review the consistent association between lean body mass catabolism with a worsening clinical outcome, increased length of hospital stay, poor recovery, and increased healthcare cost. Early enteral nutrition remains the recommended feeding route in ICU patients, although it is often unable to fully cover the nutritional needs. Parenteral nutrition, whether alone or combined with enteral nutrition, is recommended if enteral nutrition is not feasible or insufficient. Port and Apovian (pp. 184–191) discuss the particular difficulties encountered with the determination of energy requirements in obese patients. Calculating energy needs accurately is extremely problematic due to a lack of reliable prediction equations and a wide variability in body composition among the obese. Establishing protein requirements in obese, critically ill patients may be even more important, as there are data suggesting that they may be more than 2 g/kg ideal body weight per day. Andrews (pp. 192–197) reviews two hot areas in nutritional pharmacology. He elegantly discusses the evidence in favor of glutamine and selenium supplementation. The optimal route of delivery is still debated: clinical data seem to favor intravenous supplements for both nutrients. The case of intravenous selenium, which reinforces antioxidant defenses, is getting stronger in the sickest patients. Several studies analyze the glycemic control controversy issues, which started with the publication of the Leuven trial in 2001 [3], the latest installment being the Normoglycaemia in Intensive Care Evaluation-Survival Using Glucose Algorithm Regulation (NICE-SUGAR) trial [4]. The safety of tight glycemic control has been one of the major problems with this concept. In this issue, Krikorian et al. (pp. 198–204) analyze 26 protocols based on either manual calculations or computerized algorithms. Many differences and similarities are identified such as patient characteristics, target glucose level, time to achieve target glucose level, incidence of hypoglycemia, rationale for adjusting the rates of insulin infusion, and analytical methods. Several computerized protocols hold promise for safer achievement of glycemic targets, but are computers the best way? Advocates of nurse-driven protocols maintain that the latter perform best, without any definitive evidence on either side. Preiser et al. (pp. 205–210) analyze the external validity of the initial Leuven study [3], knowing that six independent, prospective, randomized controlled trials, involving 9877 patients, were unable to confirm the survival benefit reported in that pioneering trial. Among the proposed hypotheses, the case mix, the usual care, the quality of glycemic control, and the different glucose levels targeted with their risks of hypoglycemia are discussed. Finally, Scurlock et al. (pp. 211–214) discuss important arguments, which can explain the failure to repeat the initial Leuven results: interruption of two large trials [Glucontrol, Volume Substitution and Insulin Therapy in Severe Sepsis (VISEP)] for safety reasons depriving the studies from the power to achieve significant results, differences in glycemic target, early combined nutrition in the Leuven trial, and important differences in methodology (e.g., glucose measurement, insulin delivery). The difficulty in accurately replicating the original ‘proof-of-concept’ Leuven methodology may be the limiting factor for achieving the benefits gained by intensive insulin therapy. This emerging body of data suggests that it is virtually impossible in the ICU to separate intensive insulin therapy from optimal nutritional management. Unfortunately, this is not discussed openly. Is this because endocrinologists, who can provide expert consultation regarding glycemic control, are not typically formally trained in nutrition support? Or is it because physician nutrition experts, who have received formal training in nutrition support, are simply not that comfortable with aggressive intravenous insulin management, particularly in difficult, critically ill patients? This important issue has not garnered sufficient attention in the literature, training programs, clinical practice, or the research setting. Getting the glycemia right is certainly important in several categories of patients, but getting the patients adequately fed is probably at least as important [5,6]. The ASPEN/SCCM guidelines do not address this issue in their 2009 version [1]; naively, these guidelines caution programmed malnutrition by not attracting enough attention to energy delivery. Indeed, underfeeding remains a major threat to our patients: combined nutritional support appears to be a reasonable answer in this context. Nutritional management was well defined in the Leuven trial, although not traditional, with early 300 g/day of intravenous glucose: the Leuven remains the only study to have achieved proper feeding of their patients, reaching an average delivery of 19 kcal/kg/day over the first 2 weeks (Fig. 1). The majority of the other trials, including NICE-SUGAR and VISEP, have been characterized by nutritional failure, with energy deliveries around 1200–1500 kcal/day. It is, therefore, extremely difficult to compare these trials, as nutritional deficit, negative energy balance, and attendant catabolic states directly influence gluconeogenesis, intermediary metabolism, the response to tight glycemic control, and, ultimately, the outcome of critical illness [7]. The burden of underfeeding heavily biases the interpretation of the results of tight glycemic control studies. Figure 1 also shows mean energy intake per day from Villet et al.[7], a much smaller prospective study that aimed at picking up nutritional complications associated with nutritional management. The mean daily energy delivery was lower than in the Leuven trials with 16 kcal/kg/day, but higher than in the NICE-SUGAR study in which only 11 kcal/kg/day was delivered as an average, corresponding to about 50% of the patients' basal metabolic rate. The trial by Villet et al.[7] showed that the energy deficit resulting from a low feeding intake was sufficient to cause malnutrition-related complications. Committing to tight glycemic control while simultaneously underfeeding the patients does not seem prudent, given the existent data and our understanding of ICU patient physiology: this poor combination may explain the increased mortality in the tight glycemic control groups of the post-Leuven trials.Figure 1: Mean daily energy deliveries normalized for weight over the first 2 weeks‘Errare humanum est, sed perseverare diabolicum’ (Seneca) is actually what we are doing, perpetuating underfeeding of our ICU patients. Although knowing that malnutrition is deleterious, we fail to penetrate the minds of our colleagues and nursing teams with the priority constituted by its prevention. Feeding via the enteral route is difficult, indeed, but the absence of daily energy delivery monitoring contributes to the problem: not seeing the problem evolve impeaches the resolution. Computers calculate more accurately than humans, and their outputs can be customized to provide visible progression of energy balances (Fig. 2): this has been shown to improve energy delivery significantly, resulting in reduced weight loss [8]. With the pervasive influence of computerized medicine in the ICU, why cannot indirect calorimetry be a routine measurement? Why cannot these devices be built-in to ventilators? In fact, as closed-loop insulin–glucose devices are optimized, commercialized, and popularized, one can easily foresee these relevant metabolic data clearly displayed at the bedside similar to an ECG.Figure 2: Screenshot of a customized ‘page’ facilitating metabolic and nutritional monitoring, with 24-h data per columnApproaches to nutritional assessment, enteral nutrition access, parenteral nutrition access, and tight glycemic control protocols are dependent on idiosyncratic features of the ICU. Perhaps, with more research and a higher quality of data, greater standardization can result for the melding of glycemic control and nutrition support. We should also promote inclusion of training in nutrition early during the education of our young colleagues. The tasks should be better defined: the European Council stated that the absence of clear distribution of responsibilities between the various providers (physicians, nurses, and dietitians) of nutritional support was one of the major causes of underfeeding [5].