CORRESPONDENCE |
Hyperinsulinaemic normoglycaemic clamp in coronary artery surgery
* E-mail: george.carvalho{at}mail.mcgill.caEditor—We congratulate Visser and colleagues on applying glucose–insulin–potassium (GIK) therapy using a hyperinsulinaemic normoglycaemic clamp.1 Their study confirms our findings that the clamp technique is an effective, and to date the only, method of maintaining normoglycaemia in patients undergoing coronary artery bypass grafting surgery.2 In addition, they demonstrated, for the first time in this population, the attenuation of systemic inflammation with perioperative GIK therapy. More importantly, this effect was prolonged beyond the time of insulin administration providing a rationale for insulin therapy to be started before the stress of surgery and cardiopulmonary bypass (CPB).
Several issues merit further comment. It is unfortunate that when initiating the clamp the authors were unable to prevent hypoglycaemia that is, a blood glucose <3.0 mmol litre–1. Perhaps, if there had been more frequent measuring of blood glucose this complication may not have occurred. In our protocol, using an insulin dose three times that has been used in the present study, we avoided hypoglycaemia in diabetic and non-diabetic patients by measuring the blood glucose every 5 min at the start of the clamp.
We agree with the authors in stressing the importance of iatrogenic hyperglycaemia resulting from exogenous glucose administration. This is true of the metabolically unsupported control group (D5W at 30 ml h–1) and may also be true of the treatment or clamp group. Although not stated in the manuscript, we assume that the five patients excluded, due to ‘insufficient insulin therapy during CPB’, belonged to the treatment group. If so, we suspect these patients were rendered hyperglycaemic secondary to the administration of additional exogenous glucose, as for example by transfusing blood products. Packed red blood cells and fresh frozen plasma contain high concentrations of glucose that are typically >20 mmol litre–1. In anticipation of administering additional glucose, using a higher dose of insulin and therefore a higher glucose infusion rate (GIR), as proposed in our protocol, provides a larger buffer for adjusting the GIR and maintaining normoglycaemia.
Montreal, Canada
* E-mail: H.B.vanwezel{at}amc.uva.nl
Editor—We thank Drs Carvalho, Leung and Schricker for their comment on our study.1 It is true that we did not succeed in preventing hypoglycemia, (blood glucose <3 mmol litre–1) in all our patients at all times. Glucose levels between 2.2 and 3.0 mmol litre–1 were found in 7 out of 175 blood samples (4%) obtained during the first 5 h of the study. Their comment that more frequent glucose measurements taken at the start of the clamp may have prevented these individual episodes of hypoglycemia is, therefore, gratefully acknowledged. However, measuring plasma glucose at 5 min intervals, as suggested, implies the continuous use of a blood gas analyzer, equipped with a glucose electrode. This is not realistic for the routine use of GIK, as each measurement takes 
5 min. The use of a faster ‘finger-stick’ glucose measuring device, as used by Carvalho and colleagues,2 is, in our opinion, not acceptable as an alternative, due to the well-known inaccuracy of those devices. The future availability of continuous glucose measuring sensors can potentially solve those problems. None of our patients showed effects after operation, that could be attributed to the short episodes of hypoglycemia. Regarding the importance of iatrogenic hyperglycemia, there appears to be a consensus among us.1 2 In addition, we agree that glucose, which is present in packed blood cells, may contribute to elevated plasma glucose levels. However, their assumption that five GIK patients were excluded from final analysis in our study because of hyperglycemia following CPB is not correct. Those patients were excluded due to protocol violations when the insulin and glucose infusions were discontinued by mistake at the onset of CPB.
Finally, they suggested the use of much higher infusion rates of insulin, because it would lead to an increase in the infusion rate of glucose. In contrast to their data2 (which used variable infusion rates of both insulin and glucose) we actually measured plasma insulin concentrations, which were reported to be in the supra-physiological range of 600–800 pmol litre–1. We are not aware of any scientific evidence suggesting that the infusion rate of insulin described in our paper was too low. On the contrary, it cannot be excluded that their higher and varying insulin infusion rates may have led to a higher and more unpredictable rate of intracellular glucose usage, thus creating the necessity to check plasma glucose more frequently. To answer this interesting suggestion, a dose–response study in cardiac surgical patients is required. Such a study should be designed to investigate the effect of different high concentrations of insulin on the metabolic rate of glucose perioperatively.
Amsterdam, The Netherlands
References
1 Visser L, Zuurbier CJ, Hoek FJ, et al. Glucose, insulin and potassium applied as perioperative hyperinsulinaemic normoglycaemic clamp: effects on inflammatory response during coronary artery surgery. Br J Anaesth 2005; 95: 448–57
2 Carvalho G, Moore A, Qizilbash B, Lachapelle K, Schricker T. Maintenance of normoglycemia during cardiac surgery. Anesth Analg 2004; 99: 319–24
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||