Nutrition and adequate glycemic control are key management strategies associated with positive outcomes in the critically ill patient [1-4]. These components of patient management are linked since nutrition is the principal exogenous source of carbohydrate (CHO) in ventilated, tube fed patients. Nutritional formulae have the capacity to influence blood glucose control during critical illness [5]. Despite this, only a small number of studies have investigated the effect of manipulating nutritional formulae on blood glucose control in an intensive care unit (ICU) setting [6-8].

Critically ill patients experience marked changes in CHO metabolism, which contributes to hyperglycemia [9-12]. These include increased gluconeogenesis, non-insulin dependent enhanced peripheral glucose uptake, and utilization, insulin resistance and inversion of plasma glucagon to insulin ratio [11]. Hyperglycemia is prevalent in the critically ill, with up to three- quarters of patients having raised blood glucose concentrations [13]. Acute hyperglycemia is associated with various adverse outcomes including increased infectious complications, prolonged intensive care unit (ICU) and hospital stay, and an increased risk of mortality [14-16].

Historically, it was accepted as dogma that hyperglycemia was an adaptive response to stress or injury which did not require intervention [17]. This doctrine was challenged when a single center trial reported increased survival in critically ill surgical patients randomized to intensive insulin therapy aimed at maintaining blood glucose concentrations between 4.4 and 6.1mmol/L [18]. The study was pivotal to implementing strict glucose management protocols globally [19]. However, in the landmark multi-center NICE-SUGAR trial, intensive insulin therapy was linked to increased mortality compared to standard care of aiming for a blood glucose level between 6 and 10mmol/L [20]. Subsequent work has shown that it is not only hyperglycemia but also insulin-induced hypoglycemia that is strongly associated with adverse outcomes [21]. The latter has been found to be independently associated with increased mortality [22-24]. Considering that exogenous insulin affects glycemic variability [25] and, is associated with a higher risk of mortality in critically ill patients [26-28], it is appropriate to consider strategies that may facilitate reduced insulin requirements.

It has been postulated that one intervention may be substituting a more traditional exogenous nutrition with one that has a lower glycemic load and is specifically designed for hyperglycemic individuals (eg, diabetes-specific formulae [DSF]) [7]. DSF with lower CHO content has been shown to attenuate hyperglycemia and decrease glycemic variability in the non-acute setting [29-32]. However, there are limited studies evaluating the use of nutritional therapies, specifically DSF, as adjuncts to treatments for hyperglycemia in the critically ill prior to this proposed study.

Mesejo, et al. compared a high protein DSF to a standard ICU high protein formula in 50 eligible patients [7]. The aim of this open-label study was to maintain blood glucose levels between 5.5 and 11.1mmol/L through the administration of insulin as required, which is a range that varies from current practice [16,24,25]. Obese patients were excluded from participating, which has implications for the applicability of these data to countries with a greater proportion of overweight or obese patients in their ICU [33]. The investigators reported that less exogenous insulin was required to maintain acceptable glycemic control using the DSF.

De Azevedo, et al. conducted a randomized controlled trial (RCT) to assess the safety and efficacy of a CHO restrictive approach compared to intensive insulin therapy [6]. Patients allocated to the nutritional therapy group received Glucerna Select (Abbott Nutrition) and subcutaneous insulin to maintain blood glucose concentrations below 10mmol/L. The control group received an alternative formula and an insulin infusion aiming for a blood glucose target of 4.4 and 6.7mmol/L. Importantly, participants allocated to the lower target experienced a substantial increase in the frequency of hypoglycemia. These studies provide a foundation for further work investigating the role of nutrition support in the management of hyperglycemia.

In summary, glycemic control and adequate nutrition provision for critically ill patients are accepted elements of best clinical practice [34]. Hyperglycemia occurs frequently in this group, due to stress induced glucose intolerance, pre-existing diabetes mellitus and glucocorticoid therapy [18]. Management of hyperglycemia often involves exogenous insulin. Therefore, a balance of risks must be considered between hyperglycemia and treatment-induced hypoglycemia or glycemic variability [35]. The observed increase in mortality associated with insulin-induced hypoglycemia provides a strong incentive to explore managing hyperglycemia with a reduced dependency on insulin [20].

The primary aim of this study is to determine whether the administration of a DSF reduces insulin administered to critically ill patients with hyperglycemia during enteral nutrition compared to a standard liquid nutritional formula over a 48-hour period. Secondary outcomes will also be explored including the feasibility of study processes to build toward a future multi-site RCT.

A nested cohort study will also be conducted to facilitate a greater understanding of the underlying pathophysiology in critically ill patients. It will specifically seek to determine whether altered CHO and advanced glycation end product (AGE) intake, is associated with oxidative stress markers in addition to acute and chronic inflammatory biomarkers.

This is a prospective, double-blinded, single-center, two-phased, randomized controlled feasibility trial. It has been developed in accordance with the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT 2013) [36] and the Consolidated Standards for Reporting of Trials CONSORT Guidelines [37]. The trial complies with the Australian National Statement on Ethical Conduct in Human Research [38] and is currently being undertaken in the two adult ICUs at Mater Health in Brisbane, Australia. Both ICUs are run by the same intensive care specialists and protocols/ procedures are identical across sites but the ICUs differ in the health insurance status of their patients, which may be reflected in patient demographics and clinical case-mix [39].

This study has undergone full ethical review by the Mater Human Research Ethics Committee (Mater HREC; Approval HREC/14/MHS/55) and also has approval from The University of Queensland (UQ HREC; Approval 2014001353) due to the interventional nature of the research. It is also registered on the Australia New Zealand Clinical Trials Registry with registration number ANZCTR:12614000166673.

Participants will be recruited following admission to Mater Adults Hospital or Mater Private Hospital ICU. Eligibility is determined by the following inclusion criteria – a patient is receiving exclusive enteral nutrition and requires an insulin infusion. The majority of patients with two consecutive blood glucose levels > 10mmol/L will commence on an insulin infusion as per the units’ glucose management protocol. The exclusion criteria for this study include patients under the age of 18, declined consent by the patient or legally authorized representative, or if the treating physician deems participation clinically inappropriate for any patient.

Senior ICU medical staff (consultant or advanced trainee), research assistant(s) or dietitian(s) will seek informed consent from the patient preferentially where possible, or from a legally authorized representative if required where patients lack competence. Should the latter occur, patients will be approached for their consent to be followed once alert and able. A modification to the inclusion and exclusion criteria, primarily related to lowering the threshold of insulin dose used to determine eligibility, was made after commencement of this study to optimize recruitment of appropriate patients to this feasibility study.

Patient demographics including anthropometric data such as weight, height and body mass index will be collected at baseline. Additionally; admission diagnosis, acute physiology, age and chronic health evaluation (APACHE) III score, date of admission to hospital and the ICU, requirement for ventilation support, pre-existing comorbidities including pre-admission diabetic status, use of steroids and insulin at baseline will be recorded. Investigators will adopt the Patient Generated Subjective Global Assessment grading system for steroid dose which is as follows (in prednisone equivalents): low dose <10mg; moderate dose ≥10mg and <30mg and high dose ≥30mg [40].

Randomization between three arms will be performed initially 1:1:1 by a computer-generated sequence concealed within opaque sequentially numbered envelopes. Block randomization will be used to promote balanced patient numbers per arm. Envelopes will be prepared by a non-investigator and will include both the designated treatment arm (denoted as Feed A, Feed B or Feed C) as well as weight based recommended goal rates for administration of the formulae. Calculations of these prescribed rates are detailed in section 5. Once 12 patients have been recruited to the second intervention arm [Diason (Nutricia)], further recruitment to this treatment will cease, and the trial continues with the remaining interventional arm, Glucerna Select (Abbott Nutrition), and the control arm, Nutrison Protein Plus Multifibre (Nutricia).

This study will attempt blinding of the form and appearance of the formula bags from patients and study investigators. Due to the different form factors of the formulae packaging - Glucerna Select (Abbott Nutrition) is retained in a hard-plastic bottle, while Nutricia formulae are available in soft satchels, formula containers will be placed within an opaque masking bag with an opening in the bottom to allow for connection to giving sets (Figure 1). Blinding of the formulae will be carried out by a non-investigator and the concealed bags marked with the study details and treatment arm. These will be placed into corresponding storage containers for nursing staff to access.

Following consent participants will be randomized to either control, which is the unit’s standard nutritional formula, Nutrison Protein Plus Multifibre (Nutricia), or the intervention, which is a low CHO, low glycemic index DSF, Glucerna Select (Abbott Nutrition). For the purpose of the nested cohort study, a second intervention arm using a different low glycemic index DSF will be used [Diason (Nutricia)]. Refer to Table 1 for a complete nutritional breakdown of these products. Varying advanced glycation end-product content of the differing DSF necessitates the third arm as part of this study. Randomization to the second intervention arm will be ceased once a complete set of blood and urine samples are available for a minimum of 10 participants in each group, with 12 patients set as the target recruitment to account for retracted/declined consent, incomplete sample sets in the instances of extubation or discharge to the ward prior to retrieval. These samples will be collected from all patients at randomization and 48 hours later. Once this has been achieved, randomization will revert to control [Nutrison Protein Plus Multifibre (Nutricia)] and intervention [Glucerna Select (Abbott Nutrition)] groups only. Collection of urine and blood samples for biomarker assay will also be discontinued at this point.

Each envelope will include detailed instructions for clinical staff regarding nutritional formula type labeled as Feed A, B or C and a protocolized delivery rate. The rate has been determined to minimize difference in protein provision in an effort to mitigate confounding factors based on studies identifying the protein as having a protective effect [2]. Once group allocation has occurred, participants will remain on their specified study formula as long as tube feeding continues in ICU or at the discretion of the treating clinician. If the participant were to withdraw from the study, clinical care would revert to usual practice such that the formula will be changed over to what would have otherwise been used had the trial not taken place (Nutrison Protein Plus® or Diason® - Nutricia). Due to the pragmatic nature of this study, all other aspects of care will occur as per routine clinical practice. That is, rates of feeding and treatment of intolerance to enteral nutrition will be as per the unit’s standardized enteral feeding protocol.

The findings of this feasibility study will be used to refine the design of a larger multi-center trial identifying protocol strengths and limitations and importantly returning preliminary study effect estimates for sample size estimation. A more detailed budget will also be established using data from this initial study.

The pragmatic approach of negligible alterations to the pre-existing ICU policies and procedures regarding glucose management is intentional, aiming for this study to proceed with minimal disruption to standard ICU clinical work routines. If successful, a subsequent larger multi-center trial derived from the present proposed pilot study may clarify the efficacy of an intervention altering the composition of enteral nutrition in ICU patients as a standalone therapy for dysregulated glucose management in ICU patients.

Treatment of hyperglycemia, predominantly involving insulin therapy, remains controversial in the critically ill. While there is evidence to indicate diabetes specific enteral nutrition formulae are beneficial in reducing blood glucose levels in general patients with diabetes, at both a ward and community level, limited evidence supports this practice in the intensive care setting. At present, modification of the composition of enteral nutrition is rarely considered to be an efficacious mode of therapy for glucose control despite its significant contribution to total carbohydrate intake. This pilot study intends to investigate the feasibility and efficacy of diabetic specific nutrition formulations as an additional therapeutic measure in glucose management in the ICU.