In this study we demonstrate that EIC resulted in a significant reduction in mean BG at procedure end and in the first BG reading at the ICU. Additionally, EIC was associated with less hypoglycemic episodes, and less amount of insulin utilized intraoperatively. On the other hand, mean BG was higher during CPB during EIC periods compared with baseline.
To our knowledge this is the first study that uses a computerized electronic IV insulin dosing in an operating room setting in patients undergoing CPB. By so doing, a stable process of insulin dosing is created overcoming the shortcomings of human based dosing.
While implemented in the operating room, our results agree with previous studies that used electronic insulin dosing postoperatively in CPB-requiring cardiac surgeries. Chancellor et al.[8], demonstrated a significant reduction in hyperglycemic events with the use of an electronic insulin protocol coupled with an endocrinology consult service. In comparison, our study used a different cut-off for hyperglycemia (180 mg/dl vs. 200 mg/dl). While we did not formally consult an endocrinology service our EIC was created and tested by our Endocrinology collaborators. Additionally, the definition the authors protocol, was associated with higher episodes of hypoglycemia that they defined differently from us as BG< 60 mg/dl. Chancellor et al.[8], also measured clinical outcomes associated with their protocol in comparison with our study that focused on glycemic episodes at different time points.
Shelden et al[11]. used an electronic insulin protocol at the cardiovascular ICU at the University of Michigan where they demonstrated that aggressive nurse training led to an increase in conformation to target BG levels of 70-180. Yet, the protocol itself did not significantly change the incidence of hyper or hypoglycemia compared to pre-implementation. Compared to our study, Shelden et al.[11] enrolled a larger number of patients, had a different target BG and different definition of hypoglycemia. In summary, EIC has been safely and successfully implemented in the perioperative setting.
It is important to highlight that we have chosen only a sample of the baseline state during the months of October 2019. This may not be representative of the entire population at baseline. In fact, analysis of the baseline state over a longer period of time demonstrated a higher maximum average compared to our analysis. Also, we have analyzed average values at 3-time intervals intraoperatively rather than the highest values. The rationale behind this is the numerous blood draws taken during these time intervals and, more importantly, hyperglycemic value above the trigger point of EIC activation at 180 mg/dl may be too high undermining the effects of EIC by starting the protocol from too high of a blood glucose level.
We have analyzed all comers of CPB-requiring cardiac surgical patients rather than individual surgeries. These procedures vary widely in terms of duration of CPB, drugs administered on CPB and surgical practices, all factors that would have altered the blood glucose levels. Our rationale for using all comers is to test the performance of the EIC over a spectrum of blood glucose levels and to include previously overlooked yet equally affected by worse outcomes procedures such transplant and aortic surgeries.
Some of the limitations of our study, are, first, we did not analyze outcomes of hyperglycemia in our cohort such as deep sternal wound infections, lengths of hospital stay, duration on mechanical ventilation, atrial fibrillation, and mortality. The rationale for this approach is that many other factors contribute to and potentially confound the association of hyperglycemia on these outcomes. Second, we have not fully analyzed the independent association of risk factors and procedural factors on blood glucose values. This analysis is currently being performed after acquiring more data over time on our three phases. Third, we have not analyzed non CPB-requiring surgeries which we are seeing trends towards hyperglycemia due to other non-CPB related factors. This analysis is currently undertaken. Fourth, we have only analyzed blood glucose values intraoperatively short of ICU and ward values. The latter 2 locations are equally important in determining the association of hyperglycemia with adverse outcomes. Fifth, while the EIC protocol has successfully reduced end of procedure blood glucose values, it has not reduced the maximum values. For this, a third PDSA cycle has been initiated with plans to start the EIC from the beginning of the case regardless of the initial blood glucose values. We theorize that this approach will prevent overshooting of blood glucose values to levels higher than the target blood glucose of 180 mg/dl prior to starting the EIC.
In summary, The EIC has created a stable process for the intraoperative control of blood glucose in CPB-requiring cardiac surgical patients and has been successfully associated with a reduction in end of procedure average blood glucose values. The process of controlling blood glucose levels is an iterative process. More data are collected to study clinical outcomes associated with EIC implantation in our population. EIC in its ideal iteration is a generalizable process that can be intraoperatively implemented nationwide creating a new standard of intraoperative glycemic control.