The therapeutic potential of a venomous lizard: the use of glucagon-like peptide-1 analogues in the critically ill

Glucagon-like peptide-1 (GLP-1), a principal mediator of the postprandial insulinotropic response in health, has a half-life of minutes. The saliva of the Gila monster contains exendin-4, a structural analogue of human GLP-1, but with a much longer half-life. A synthetic preparation of exendin-4, exenatide, is suitable for human use and effectively lowers glucose in ambulant type 2 diabetic patients. When compared with insulin, exenatide therapy is associated with a reduction in hypoglycaemic episodes and postprandial glycaemic excursions in this group. Accordingly, GLP-1 analogues are appealing therapies for hyperglycaemia in the critically ill patient and warrant further study.

exogenous GLP-1 are glucose dependent, such that even pharma co logical doses of GLP-1 are most unlikely to cause hypo glycaemia.
Native GLP-1 is not used as a glucose-lowering agent in ambulant type 2 diabetic patients because of the necessity for continuous administration [4]. Endogenous GLP-1 is metabolised rapidly to its so-called inactive metabolite by the ubiquitous enzyme dipeptyl-peptidase-4 (DPP-4), the half-life of native GLP-1 being ~1 to 2 minutes [5]. While this factor essentially precludes the use of exogenous GLP-1 in ambulant patients, rapid metabolism of an agent may be a desirable pharmacokinetic property in the hospitalised patient. Accordingly, our group and others have explored the eff ects of exogenous GLP-1 in a variety of hospitalised, adult patients [6][7][8][9][10].
Usin g a prospective, randomised, open-label, design, Mecott and colleagues compared an intervention (exenatide ± insulin) with a control (intensive insulin therapy) i n patients receiving small intestinal nutrient. Th e authors report that the intervention decreased insulin requirements while comparable glycaemic control was achieved. Th e therapeutic use of GLP-1 in critically ill patients is inherently attractive as it does not substantially increase the risk of hypoglycaemia [11]. In addition, GLP-1 therapies show reduced glycaemic variability when compared with insulin in ambulant diabetic patients [11].
Strengths of the study are the evaluation of a GLP-1based therapy in a new population (paediatric) that has suff ered a specifi c insult (burns), and the prolonged duration of evaluation (up to 28 days). Finally, and perhaps most signifi cantly, the investigators have elected to use a GLP-1 analogue (exenatide) rather than the synthetic peptide. Th e major advantages of exenatide when compared with GLP-1 are that it can be administered intermittently, as well as its cheaper price and ready availability ( Table 1).
Limitations of the study, which are largely acknowledged by the authors, include its open-label design and the relatively small number of subjects studied, with the consequent potential for type 2 errors. Th e authors reported that the use of the GLP-1 analogue failed to reduce hypoglycaemic episodes and glycaemic variability. However, the lack of eff ect on variability and hypoglycaemia may refl ect administration of exenatide ± insulin, rather than using exenatide as a single agent. Plasma exenatide concentrations were not reported, and it should be recognised that concentrations may not be predictable in these subjects -since, even in healthy subjects, exenatide has a biological half-life of ~3 hours and large fl uctuations in plasma levels occur with twicedaily subcutaneous administration. Furthermore, given the limited infor mation on the mechanisms of glucoselowering in the critically ill patient, measurement of insulin and/or C-peptide, as well as glucagon, would have been valuable.
Exendin-4 was isolated originally from the saliva of the Gila monster (Heloderma suspectum), a slow-moving venomous lizard native to the United States and Mexico. Th e Gila monster eats only 5 to 10 times per year and a meal causes a substantial postprandial increase in plasma exendin-4 concentrations [12]. Exendin-4 shares ~50% amino acid sequence identity with human GLP-1 and binds to the pancreatic GLP-1 receptor in vitro. Importantly, exendin-4 is resistant to DPP-4 inactivation, thereby having a prolonged duration of action [12]. Exenatide, a synthetic form of exendin-4, has been shown to reduce fasting and postprandial glucose via glucosedependent stimulation of insulin and suppression of glucagon secretion, as well as slowing gastric emptying [13]. Th e latter is probably the dominant mechanism to account for glucose lowering after oral or intragastric meals in healthy subjects, type 2 diabetic patients, and those critically ill patients in whom gastric emptying is normal [10,13]. GLP-1 analogues with half-lives between 12 hours and 3 to 4 days have been recently developed. Th ese analogues have less variation in plasma concentrations after once-daily (for example, liraglutide) or weekly (sustained release exenatide) administration than twice a day dosing (exenatide). Oral drugs that inactivate the DPP-4 enzyme (DPP-4 inhibitors) -thereby attenuating metabolism and increasing the availability of endogenous GLP-1 and GIP -have also entered the clinical domain (Table 1).
In the study of Mecott and colleagues, exenetide was well tolerated; but given the adverse eff ects associated with its use in ambulant diabetic patients, ongoing vigilance is warranted. Th ese adverse eff ects include nausea and vomiting (usually transient) [14], as well as modest weight loss that occurs, and is frequently desirable, in ambulant type 2 diabetic patients [14], but may be detrimental in the critically ill patient. Th e potential association between GLP-1 analogues and pancreatitis remains uncertain [14]. Lastly, the investigators did not measure anti-exenatide antibodies or plasma calcitonin concentrations, but the signifi cance of anti-exenatide antibiodies or mild increases in calcitonin concentrations is unclear [14].
Albeit preliminary, the present study represents part of a growing interest in the use of incretins, or incretin mimetics, for the management of hyperglycaemia in hospitalised patients [15]. Th e desirable blood glucose range in this group remains contentious, but, as with recent studies in ambulant type 2 diabetic patients, there seems to be minimal, or no, advantage in targeting glycaemia at the lower end of the fasting normal range. Rather, the latter may well be deleterious [16]. Because the glucose-lowering eff ect of GLP-1 is glucose dependent, there is likely to be a threshold -perhaps Table 1 about 7 to 8 mmol/l -beyond which further reductions will only occur if exogenous insulin is co-administered. In addition to identifying the optimal glycaemic range, future studies should focus on the following: which of the incretin agents is most useful in the critically ill patient; whether these agents should be used in combination with insulin, or as single-agent therapy; and identifi cation of the patient group most likely to benefi t from administration of incretin mimetics.