In this study we found that cardiac surgical patients who received two doses of dexamethasone displayed higher blood glucose levels between 6 and 24 hours after ICU admission than patients who received one dose of dexamethasone.
High dose dexamethasone is a controversial therapy in cardiac surgery patients. Both studies that favor the use of corticosteroids [5
] or studies that show some benefit [6
] as well as studies that show no benefit [7
] have been published. The effect of steroids on glucose control may be of interest, because high glucose levels are common and have been shown to be independent predictors of adverse outcome during cardiac surgery [11
]. Treatment of hyperglycemia by continuous insulin infusion has recently been shown to improve outcome [12
]. The difference of 1.5 mmol/L in glucose levels we found in multivariable analysis is clinically relevant considering a post-hoc analysis found that the odds ratio for ICU mortality for every increase in glucose level of 1.1 mmol/L is 1.3 [14
]. Our study was not powered to detect a difference in mortality.
While the merit of steroids for cardiopulmonary bypass may be disputed, steroids are administered for many other indications in other critically ill patients. For instance, lower dose corticosteroids may be used to prevent atrial fibrillation after cardiac surgery [21
]. For other patient categories, effects on glucose control have recently been shown for hydrocortisone in patients with septic shock [22
], or for dexamethasone against nausea in abdominal surgery patients [23
A number of limitations of our study must be mentioned. First, we had not implemented a tight glucose control protocol at the thoracic ICU at the time of this study. Therefore, overall glucose control was not comparable to the levels achieved in intervention studies, reflected by the HGI which indicates mean glucose levels of 9.5 and 10.7 mmol/L for the groups respectively. We think this does not change our conclusion that dexamethasone induces hyperglycemia, as we hypothesize that under a tight glucose control protocol, we would have found that the 1D group would require less interventions and lower doses of insulin than the 2D group to achieve the same glucose targets. Second, although data collection was prospective, the study as a whole must be regarded as a retrospective study. Bias could have been introduced in our data in a number of ways. Inherent to the before-after design the patients were not randomly allocated to the two dosing groups. A very important variable, insulin infusion, was found to be significantly different between the two groups. The multivariable analysis incorporating this and other possible confounding factors confirmed the highly significant relation of glucose levels with the administration of one or two dexamethasone doses. This analysis may still overestimate the difference caused by dexamethasone due to unknown confounding factors and due to non-linear effects of the included variables. Not all cardiac surgery centers use dexamethasone as routine treatment in all cardiac surgery patients. This may be a limitation to the relevance of our study. However, it is unlikely that the hyperglycemic effect we measured in this study is unique to dexamethasone or cardiac surgery patients. Our findings may be extrapolated to the majority of patients receiving corticosteroids in the intensive care setting.
Although we did not study the effect of giving dexamethasone compared to no dexamethasone is this study, the first dose of dexamethasone also induced considerable insulin resistance as we have observed in a previous study [18
]. Withholding of the dose administrated upon induction of anesthesia could further improve glucose control around the time of ICU admission. This quantitative study shows that one should be prepared to administer higher doses of insulin when using dexamethasone in cardiac surgery patients. In three ICUs in our hospital, we have implemented a computerized glucose control system. We designed this system so that patients who receive steroids both get higher initial insulin doses, and are checked more frequently to promptly detect hyperglycemia [24
]. Future research on the value of corticosteroids during cardiac surgery or for other indications in the ICU should pay very close attention to the hyperglycemic effects of these drugs. If glucose control is not performed adequately, potential positive effects of steroids could be offset by iatrogenic hyperglycemia.
In summary, the administration of dexamethasone exerts a considerable hyperglycemic effect in cardiac surgery patients, interfering with glucose control.