Hyperlactataemia is a useful surrogate marker representing anaerobic metabolism resulting from an imbalance between oxygen supply and demand in the tissues.1 It is frequently encountered in cardiac surgery requiring cardiopulmonary bypass (CPB) and has been demonstrated to be associated with adverse outcomes.2–5 Risk factors for hyperlactataemia in cardiac surgery include complex surgery, longer duration of CPB, haemodynamic instability, use of vasoconstrictors and hyperglycaemia.2–4 Inherent to the use of CPB, various factors including insufficient pump flow, excessive haemodilution and ischaemia-reperfusion injury are known to contribute to increased myocardial and splanchnic lactate production and compromised hepato-splanchnic lactate extraction.3
Through its principal metabolic mechanism switching the myocardium to use glucose as the primary fuel source instead of free fatty acids, glucose–insulin–potassium (GIK) solution has been used for decades as an adjuvant for myocardial protection in cardiac surgery.6–12 In addition to its beneficial metabolic influences, insulin in GIK has been reported to produce systemic vasodilatation and improve blood flow to major organs including the liver and intestines under diverse conditions.13–16 Thus, on the basis of its reported beneficial effects, GIK may be able to influence lactate production through attenuating splanchnic as well as myocardial lactate production and improving hepatic lactate extraction during CPB. In addition, improvement in myocardial performance with GIK may favour haemodynamic stability and proper oxygen supply to the tissues, resulting in a diminished incidence of hyperlactataemia during the perioperative period in the cardiac surgical setting. However, most studies evaluating the effects of GIK in cardiac surgery have focused on myocardial performance or myocardial injury without strict glycaemic control,17–20 whereas no comprehensive data exist regarding its influence on systemic lactate concentration and outcome.
This prospective, randomised and controlled trial was designed to investigate the effect of intraoperative GIK infusion under tight glycaemic control on the incidence of hyperlactataemia and on postoperative outcome in patients at risk of developing hyperlactataemia in association with valvular heart surgery.
Materials and methods
After approval by the Institutional Review Board (Institutional Review Board of Yonsei University Severance Hospital chaired by Chung Mo Nam, PhD; 50 Yonsei-ro, Seodaemun-gu, Seoul, South Korea; protocol number, 4-2012-0347; approved on 11 July 2012) and after written informed consent had been obtained, consecutive patients undergoing cardiac surgery under CPB between March 2013 and October 2013 at Severance Hospital were screened and 106 patients with at least two of the following risk factors for hyperlactataemia were enrolled: congestive heart failure; re-do surgery; complex surgery (combined mitral valve surgery/tricuspid annuloplasty or Maze procedure, double-valve surgery, combined aortic valve replacement and ascending aorta replacement, Bentall's operation, combined valve/coronary artery bypass graft surgery, graft replacement of aorta); expected duration of CPB longer than 2 h; and preoperative left ventricular ejection fraction (LVEF) lower than 40%.
Patients with the following profiles were excluded: emergency surgery; insulin-dependent diabetes mellitus; preoperative haemodynamic instability requiring mechanical or pharmacological cardiopulmonary support; baseline lactate concentration more than 2 mmol l−1; and liver dysfunction (glutamyl transferase concentrations more than double upper normal limits). Patients were randomly assigned either to the GIK group (n = 53) or control group (n = 53) according to random numbers produced by a computer.
After induction of general anaesthesia, patients in the GIK group received insulin 0.1 IU kg−1 h−1 and an infusion of a mixture of 30% glucose and 80 mmol l−1 of potassium at a rate of 0.5 ml kg−1 h−1 until the completion of the operation. The infusion rate of the glucose and potassium mixture was adjusted to maintain the blood glucose concentration between 4.4 and 11.1 mmol l−1 (80 to 200 mg dl−1) and potassium concentration between 3.5 and 5.5 mmol l−1 according to the results of serially performed arterial blood gas analyses. In the control group, patients received equivalent volumes of 0.9% saline, and intravenous boluses of insulin (1 IU per 2.2 mmol l−1 increase in blood glucose concentration) were used to maintain the blood glucose concentration within the same predefined target range. Thus, the anaesthesia provider could not be blinded, but staff in the ICU were blinded to the patients’ group and perioperative profiles.
Patients were provided with standardised perioperative care according to institutional guidelines. After application of standard monitors including five-lead electrocardiography, invasive arterial blood pressure monitor and pulmonary artery catheter, anaesthesia was induced with midazolam, sufentanil and rocuronium, and maintained with sufentanil, sevoflurane and vecuronium to keep the bispectral index score between 40 and 60. Transoesophageal echocardiography was performed during the operation as needed. Mechanical ventilation was applied to maintain the end-tidal carbon dioxide concentration between 4.0 and 4.7 kPa, with an inspired oxygen fraction (FIO2) of 0.4 and 5 cmH2O of positive end-expiratory pressure. CPB was conducted using a membrane oxygenator at an FIO2 of 0.6 and nonpulsatile pump flow at a rate of 2.0 to 2.4 l min−1 m−2 under mild hypothermia (oesophageal temperature between 32°C and 34°C) using α-stat management. Mean arterial pressure was maintained between 50 and 80 mmHg during the perioperative period with norepinephrine, vasopressin or sodium nitroprusside as needed. Vasopressors were used in a stepwise additive fashion. The primary vasopressor was noradrenaline up to 0.5 μg kg−1 min−1, and vasopressin (up to 4 IU h−1) was added if the desired mean arterial pressure could not be achieved. If the cardiac index was below 2.0 l min−1 m−2 despite volume augmentation, milrinone was used as the first-line inotrope (0.3 to 0.7 μg kg−1 min−1). The haematocrit was maintained above 20% during CPB and above 25% at other times. All patients were transferred to the ICU after surgery.
The primary endpoint was the incidence of hyperlactataemia (lactate concentration ≥4 mmol l−1) during the operation and until 24 h after the operation. Absolute lactate values were also compared between the groups. Secondarily, a composite of postoperative morbidity endpoints and operative mortality were compared between the groups. Assessed morbidity endpoints included permanent stroke, renal failure [new requirement of dialysis or an increase of the serum creatinine concentration by >177 μmol l−1 (2 mg dl−1) or to double the baseline concentration], prolonged mechanical ventilation for more than 24 h, deep sternal wound infection, reoperation for any reason and myocardial infarction (newly developed Q wave or creatine kinase-MB elevated to more than five times the upper normal limit). Operative mortality was defined as death during the same hospitalisation period.
Patients’ characteristics and related data included patients’ demographics, comorbidities, LVEF, EuroSCORE, medications, type of operation, durations of CPB, aortic cross-clamp (ACC) and operation and laboratory data. The number of patients exhibiting hypoglycaemia (blood sugar concentration <4.4 mmol l−1) or hyperglycaemia (blood sugar concentration >11.1 mmol l−1) at least once throughout the study period and the amount of insulin administered during the entire study period were recorded. Glycaemic variability and time weight average (TWA) of glucose concentration were assessed from the glucose concentrations obtained during and 24 h after surgery. Glycaemic variability was defined as standard deviation divided by mean of glucose concentration. Time weight average was calculated as area under the curve for the glucose measurements divided by the time of each measurement with the following equation.
where Δt = time interval between each measurement (min); n = time interval number; N = total number of time intervals; c = glucose concentration at the end of each time interval.
Intraoperative and postoperative parameters for the 24-h period after operation included haemodynamic and laboratory data, fluid balance and vasopressor/inotrope requirements. Haemodynamic data included heart rate, mean arterial blood pressure, mean pulmonary arterial pressure, central venous pressure, cardiac index and mixed venous oxygen saturation (SvO2). Evaluated laboratory data were results of arterial blood gas analysis including glucose and potassium concentrations. Haemodynamic and laboratory data were recorded before induction of anaesthesia (T0), 10 min after induction (T1), 5 min after ACC removal (T2), 10 min after CPB weaning (T3), completion of operation (T4), ICU admission (T5), 3 h after ICU admission (T6) and 24 h after operation (T7). For fluid balance, the volumes of infused crystalloid, colloid and blood components transfused, urine output during the operation and ICU stay were recorded. Pharmacological support including the use of noradrenaline, vasopressin, milrinone and/or dobutamine were recorded throughout the study period. Lengths of ICU and hospital stay were also recorded.
The primary endpoint was the incidence of perioperative hyperlactataemia. In a retrospective review of institutional data, the incidence of perioperative hyperlactataemia in this subset of patients was approximately 60%. Under the assumption that a 50% reduction in the incidence of hyperlactataemia should be significant, a sample size of 48 for each group was needed at 80% power and α-level of 0.05. Considering a dropout of 10%, a sample size of 53 patients per group was determined. Statistical analysis was performed with SAS version 9.2 (SAS Institute Inc., North Carolina, USA). Variables are presented as either mean ± SD or numbers, as appropriate. For continuous variables, either χ2 or Fisher's exact test was performed. The Student t-test was used for categorical variables. For repeatedly measured variables, either linear mixed model or logistic regression with a generalised estimating equation method was used as appropriate under the assumption that the correlations between measurement points were identical.
The study was performed successfully in all 106 patients without exclusion (Fig. 1). Baseline characteristics including operative and anaesthetic data were similar in the two groups (Table 1).
Perioperative haemodynamics and related data are summarised in Table 2. SvO2 (P = 0.028) at T4 was significantly higher in the GIK group than in the control group.
The incidence of hyperlactataemia was the same in each group (32/53 patients, P > 0.999). The numbers of patients exhibiting hypoglycaemia (four vs. nine patients in the control and GIK group, respectively, P = 0.184) or hyperglycaemia (37 vs. 42 patients in the control and GIK group, respectively, P = 0.652) were not significantly different between the groups. The total amounts of insulin administered during the study period were 9 ± 11 and 26 ± 12 IU in the control and GIK groups; respectively. Perioperative glycaemic variabilities were similar between the groups (0.39 vs. 0.40 intraoperatively and 0.29 vs. 0.26 postoperatively in the control and GIK groups; P = 0.724 and 0.262 respectively). Time weight averages during and after surgery were also very similar in each group (8.2 vs. 8.1 mmol l−1 intraoperatively and 8.3 vs. 8.6 mmol l−1 postoperatively in the control and GIK groups; P = 0.909 and 0.502 respectively). Lactate concentrations at T2 (P = 0.010) and T3 (P = 0.013) were significantly higher in the GIK group than in the control group. Glucose concentration at T5 was significantly lower in the GIK group than in the control group (P = 0.017). Serum potassium concentrations at T3-T5 were significantly lower in the GIK group than in the control group. Preoperative and postoperative serum creatinine and CK-MB concentrations were similar in each group (Table 3).
Perioperative fluid balance and pharmacological support data were similar in both groups (Table 4).
The number of patients developing predefined postoperative morbidity endpoints and mortality were not significantly different between the groups (Table 5).
In this prospective, randomised control and controlled trial, intraoperative infusion of GIK solution accompanied by tight glycaemic control did not exert any beneficial influence on the incidence of perioperative hyperlactataemia or a composite of morbidity endpoints.
Blood lactate concentration has long been regarded as a marker of an imbalance between tissue oxygen supply and demand.1 In cardiac surgery using CPB, well known risk factors for hyperlactataemia such as low LVEF, complex or re-do surgery, low haemoglobin concentration, haemodynamic instability and prolonged duration of CPB2–4 implied that lower oxygen supply to the tissues than their demands led to anaerobic metabolism.3 Indeed, strong correlations between hyperlactataemia and postoperative morbidity and mortality have been repeatedly confirmed in the cardiac surgical setting,2–4 whereas preventive measures against hyperlactataemia have not been fully investigated.
Glucose–insulin–potassium infusion has been shown to protect the myocardium and improve myocardial performance against ischaemia-reperfusion injury through metabolic and nonmetabolic mechanisms.7–12 Its primary metabolic mechanism is to increase glucose concentration and decrease free fatty acid concentration in the myocardium, which renders the heart more energy-efficient under ischaemic conditions.21 In addition, insulin directly activates cell-surviving pathways related to tyrosine kinase and the phosphatidylinositol-3-kinase-AKT-endothelial nitric oxide synthase signalling pathway, which improve myocardial cell viability and reduce apoptosis.22,23 Insulin also has direct inotropic and vasodilating effects leading to an increased cardiac output.24,25 These effects could result in enhanced organ blood flow and improvements in splanchnic, hepatic and renal blood flows with GIK infusion have been observed in experimental and clinical conditions.13,15,16 Moreover, there was a report that insulin augmented oxygenation for central mixed and hepatic venous blood in cardiac surgery.14 Because the removal of lactate in the liver has a major impact on serum lactate concentration, improved liver blood flow would augment lactate removal. It has also been shown that GIK infusion reduced lactate release during cardioplegic cardiac arrest in patients undergoing coronary revascularisation under CPB.12 Thus, the mechanisms of action of GIK would be expected to reduce lactate production and increase its clearance, although evidence regarding its influence on overall systemic lactate concentrations in the cardiac surgical setting is lacking.
In previous studies addressing the influence of GIK, heterogeneous protocols have been applied. Of importance, the beneficial effects of GIK were known to be more prevalent when insulin was given in high dose during reperfusion in high-risk patients with minimised disturbances in glucose homeostasis.12,24,26,27 Thus, to maximise the effects of GIK, we infused a dose of insulin known to be sufficient to suppress circulating concentrations of free fatty acids (≥5 IU h−1) consistently throughout the operation in patients undergoing complex valvular heart surgery requiring prolonged CPB.26 To maintain normal blood glucose and potassium concentrations, the infusion of glucose and potassium was independent of the insulin infusion for practical rate adjustment.
However, GIK infusion did not influence the incidence of hyperlactataemia or the increase in absolute blood lactate concentration in this study. Glucose–insulin–potassium infusion did not result in significant improvements in haemodynamic parameters or outcome. The results of the current trial are inconsistent with the results of a recent study reporting a significant decrease in the incidence of low cardiac output syndrome in patients undergoing aortic valve replacement.7 We also found no beneficial influence of GIK in terms of myocardial enzyme release, although a recent meta-analysis involving adult cardiac surgical patients reported reduced myocardial injury and improved haemodynamic performance related to the use of GIK.28 Of note, serum lactate concentrations showed trends towards being higher in the GIK group in the current trial, which began after reperfusion.
A possible explanation for the results of the current study is an increase in lactate production merely by increased glycolysis and glucose metabolism imposed by GIK administration.29 Insulin induces glucose uptake in the absence of ischaemia when oxidative phosphorylation is fully functioning. Thus, glycolysis rates would be expected to increase and the resulting superfluous pyruvate would be converted to lactate; the clinical significance is unknown because the increase in lactate is not associated with ischaemic injury and acidosis.30 Of concern, however, would be the increase in ’anaerobic’ glycolysis by GIK administration during low-flow ischaemia or reperfusion and the resultant increase in lactate production and intracellular acidosis.31 The consequences of this potential drawback are still a matter of debate, but none of the meta-analyses of GIK studies has revealed any negative results to date.32,33
In the current trial, enrolled patients had two or more risk factors for hyperlactataemia and the overall incidence of hyperlactataemia was 60% as we expected, which was higher than the results of previous studies (approximately 20%).2–5 Patients developing hyperlactataemia showed higher incidences of renal failure and reoperation and longer durations of ventilator care, ICU and hospital stay regardless of the group in the current trial (data not shown). These results implicate the value of lactate concentration as a predictor of poor prognosis. Nonetheless, morbidity endpoints between the groups were not different and patients in the GIK group showed a trend towards a lower incidence of developing renal failure. However, it is beyond the scope of the current trial to draw conclusions regarding the influence of GIK on outcome, as it was not powered to address this matter.
Another plausible explanation would be that although hyperglycaemia is considered to abolish the beneficial effect of GIK,34 there was a theory that increased osmolarity of the extracellular space caused by hyperglycaemia with GIK infusion could reduce cellular oedema during ischaemia.24,35 Indeed, hyperglycaemia frequently developed in the GIK group even in a study reporting positive results.7 However, blood glucose concentration was tightly controlled in the current study, resulting in a lower concentration in the GIK group on ICU admission. As noted above, tight glycaemic control could increase the risk of undetected hypoglycaemia,36 which might have offset any potential beneficial effect of GIK in this study. In this study, however, perioperative glycaemic variability and TWA were reasonably low and similar in both groups. In addition, none of the patients developed moderate to severe hypoglycaemia (<3.3 mmol l−1). Therefore, it is less likely that glucose control affected the results of the present study. In addition, 27 out of 53 patients in the control group received insulin during the operation to control hyperglycaemia. Considering that some protocols using single-bolus or low-dose insulin improved cardiac output,26,27,37,38 the use of insulin in the control group might have mitigated the differences in outcome between the groups. Lastly, insulin is known to activate plasminogen activator inhibitor 1, a marker representing the inflammatory state, which is associated with increased serum lactate concentration after cardiac surgery on CPB.39,40
This study has a number of limitations. First, although the anaesthetic caregivers were not aware of this study, they could not be blinded due to the difference in glycaemic control protocols between the groups. Second, it might have been more beneficial to continue GIK infusion after surgery as previously suggested,7,20,24 whereas GIK infusion was discontinued when the patient was transferred to the ICU in the current trial, usually 2 to 3 h after reperfusion. Although the dose we used is known to be enough to suppress free fatty acids in serum, we did not measure the serum free fatty acid concentrations to prove otherwise. Also, the infusion duration might not have been long enough to convey any systemic influence.24 Lastly, lactic acid production is also increased by nondysoxic stimulants, such as inflammation and catecholamines.41 Although our study was a randomised, controlled trial and both groups were identical in terms of durations of CPB and ACC, and use of vasoactive drugs, the influence of the type B lactic acidosis was not separately assessed and may act as a confounder.
In conclusion, GIK infusion during surgery with tight glycaemic control did not reduce the incidence of perioperative hyperlactataemia in high-risk patients undergoing complex valvular heart surgery. In addition, it was not associated with definite improvements in haemodynamics or myocardial enzyme release.
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship: none.
Conflicts of interest: none.
Presentation: preliminary data from this study were presented as a poster at the 90th Annual Scientific meeting of the Korean Society of Anesthesiologists, November 2013, Jungsun, Korea.
1. Kapoor P, Mandal B, Chowdhury U, et al. Changes in myocardial lactate, pyruvate and lactate-pyruvate ratio during cardiopulmonary bypass for elective adult cardiac surgery: early indicator of morbidity. J Anaesthesiol Clin Pharmacol
2. Demers P, Elkouri S, Martineau R, et al. Outcome with high blood lactate levels during cardiopulmonary bypass in adult cardiac operation. Ann Thorac Surg
3. Ranucci M, De Toffol B, Isgro G, et al. Hyperlactatemia during cardiopulmonary bypass: determinants and impact on postoperative outcome. Crit Care
4. Maillet JM, Le Besnerais P, Cantoni M, et al. Frequency, risk factors, and outcome of hyperlactatemia after cardiac surgery. Chest
5. Kogan A, Preisman S, Bar A, et al. The impact of hyperlactatemia on postoperative outcome after adult cardiac surgery. J Anesth
6. Metabolic support for the postischaemic heart. Lancet
7. Howell NJ, Ashrafian H, Drury NE, et al. Glucose-insulin-potassium reduces the incidence of low cardiac output episodes after aortic valve replacement for aortic stenosis in patients with left ventricular hypertrophy: results from the Hypertrophy, Insulin, Glucose, and Electrolytes (HINGE) trial. Circulation
8. Rowe JW, Young JB, Minaker KL, et al. Effect of insulin and glucose infusions on sympathetic nervous system activity in normal man. Diabetes
9. Vlasselaers D. Glucose-insulin-potassium: much more than enriched myocardial fuel. Circulation
10. Oates A, Nubani R, Smiley J, et al. Myocardial protection of insulin and potassium in a porcine ischemia-reperfusion model. Surgery
11. Revelly JP, Tappy L, Martinez A, et al. Lactate and glucose metabolism in severe sepsis and cardiogenic shock. Crit Care Med
12. Kjellman UW, Bjork K, Dahlin A, et al. Insulin (GIK) improves myocardial metabolism in patients during blood cardioplegia. Scand Cardiovasc J
13. Bronsveld W, van Lambalgen AA, van den Bos GC, et al. Regional blood flow and metabolism in canine endotoxin shock before, during, and after infusion of glucose-insulin-potassium (GIK). Circ Shock
14. Lindholm L, Bengtsson A, Hansdottir V, et al. Insulin (GIK) improves central mixed and hepatic venous oxygenation in clinical cardiac surgery. Scand Cardiovasc J
15. Jeppsson A, Ekroth R, Kirno K, et al. Insulin and amino acid infusion after cardiac operations: effects on systemic and renal perfusion. J Thorac Cardiovasc Surg
16. Lang CH, Alteveer RJ. Effects of glucose-insulin-potassium on intestinal hemodynamics and substrate utilization during endotoxemia. Am J Physiol
17. Besogul Y, Tunerir B, Aslan R, et al. Clinical, biochemical and histochemical assessment of pretreatment with glucose-insulin-potassium for patients undergoing mitral valve replacement in the third and fourth functional groups of the New York Heart Association. Cardiovasc Surg
18. Vinten-Johansen J, Buckberg GD, Okamoto F, et al. Superiority of surgical versus medical reperfusion after regional ischemia. J Thorac Cardiovasc Surg
19. Lazar HL, Zhang X, Rivers S, et al. Limiting ischemic myocardial damage using glucose-insulin-potassium solutions. Ann Thorac Surg
20. Shim JK, Yang SY, Yoo YC, et al. Myocardial protection by glucose-insulin-potassium in acute coronary syndrome patients undergoing urgent multivessel off-pump coronary artery bypass surgery. Br J Anaesth
21. Lopaschuk GD, Wambolt RB, Barr RL. An imbalance between glycolysis and glucose oxidation is a possible explanation for the detrimental effects of high levels of fatty acids during aerobic reperfusion of ischemic hearts. J Pharmacol Exp Ther
22. Jonassen AK, Brar BK, Mjos OD, et al. Insulin administered at reoxygenation exerts a cardioprotective effect in myocytes by a possible antiapoptotic mechanism. J Mol Cell Cardiol
23. Sack MN, Yellon DM. Insulin therapy as an adjunct to reperfusion after acute coronary ischemia: a proposed direct myocardial cell survival effect independent of metabolic modulation. J Am Coll Cardiol
24. Doenst T, Bothe W, Beyersdorf F. Therapy with insulin in cardiac surgery: controversies and possible solutions. Ann Thorac Surg
25. Doenst T, Richwine RT, Bray MS, et al. Insulin improves functional and metabolic recovery of reperfused working rat heart. Ann Thorac Surg
26. McDaniel HG, Papapietro SE, Rogers WJ, et al. Glucose-insulin-potassium induced alterations in individual plasma free fatty acids in patients with acute myocardial infarction. Am Heart J
27. Gao F, Gao E, Yue TL, et al. Nitric oxide mediates the antiapoptotic effect of insulin in myocardial ischemia-reperfusion: the roles of PI3-kinase, Akt, and endothelial nitric oxide synthase phosphorylation. Circulation
28. Fan Y, Zhang AM, Xiao YB, et al. Glucose-insulin-potassium therapy in adult patients undergoing cardiac surgery: a meta-analysis. Eur J Cardiothorac Surg
29. Dimitriadis G, Parry-Billings M, Bevan S, et al. The effects of insulin on transport and metabolism of glucose in skeletal muscle from hyperthyroid and hypothyroid rats. Eur J Clin Invest
30. Sahuquillo J, Merino MA, Sanchez-Guerrero A, et al. Lactate and the lactate-to-pyruvate molar ratio cannot be used as independent biomarkers for monitoring brain energetic metabolism: a microdialysis study in patients with traumatic brain injuries. PLoS One
31. Neely JR, Grotyohann LW. Role of glycolytic products in damage to ischemic myocardium. Dissociation of adenosine triphosphate levels and recovery of function of reperfused ischemic hearts. Circ Res
32. Schipke JD, Friebe R, Gams E. Forty years of glucose-insulin-potassium (GIK) in cardiac surgery: a review of randomized, controlled trials. Eur J Cardiothorac Surg
33. Bothe W, Olschewski M, Beyersdorf F, Doenst T. Glucose-insulin-potassium in cardiac surgery: a meta-analysis. Ann Thorac Surg
34. LaDisa JF Jr, Krolikowski JG, Pagel PS, et al. Cardioprotection by glucose-insulin-potassium: dependence on KATP channel opening and blood glucose concentration before ischemia. Am J Physiol Heart Circ Physiol
35. Okamoto F, Allen BS, Buckberg GD, et al. Reperfusate composition: interaction of marked hyperglycemia and marked hyperosmolarity in allowing immediate contractile recovery after four hours of regional ischemia. J Thorac Cardiovasc Surg
36. Preiser JC, Devos P, Ruiz-Santana S, et al. A prospective randomised multicentre controlled trial on tight glucose control by intensive insulin therapy in adult intensive care units: the Glucontrol study. Intensive Care Med
37. Svensson S, Berglin E, Ekroth R, et al. Haemodynamic effects of a single large dose of insulin in open heart surgery. Cardiovasc Res
38. Diaz R, Paolasso EA, Piegas LS, et al. Metabolic modulation of acute myocardial infarction. The ECLA (Estudios Cardiologicos Latinoamerica) Collaborative Group. Circulation
39. Dixon B, Santamaria JD, Campbell DJ. Plasminogen activator inhibitor activity is associated with raised lactate levels after cardiac surgery with cardiopulmonary bypass. Crit Care Med
40. Nordt TK, Sawa H, Fujii S, Sobel BE. Induction of plasminogen activator inhibitor type-1 (PAI-1) by proinsulin and insulin in vivo. Circulation
© 2015 European Society of Anaesthesiology
41. James JH, Luchette FA, McCarter FD, Fischer JE. Lactate is an unreliable indicator of tissue hypoxia in injury or sepsis. Lancet