Letters to the Editor: Letter to the Editor
We thank Sessler1 for his interest in our recent article on the utility of intraoperative ST segment values for predicting postoperative troponin elevation.2 It is an honor to receive these comments from an expert in the field of myocardial injury after noncardiac surgery, and we are grateful to have the opportunity to reply.
In response to his first point, we concur that our study design is better described as a retrospective cohort study. The next topic, which pertains to the possible impact that selection bias had on our results, requires more discussion. We acknowledged in our article that selection bias may have influenced the analysis because the decision to measure postoperative troponin levels was made using clinical judgment. As pointed out by Sessler, ST segment depression is commonly used to justify measuring postoperative troponin levels. On the basis of this, if our study only examined ST segment depression, it would be expected that the association between intraoperative ST segment values and postoperative troponin elevation would be overestimated. However, our study included other characteristics of ST segments and demonstrated that these other features may be more important.
This highlights one of the most intriguing finding of this study, namely, that the variability of ST segments had a greater impact than ST segment depression with increased standard deviation being associated with decreased risk of postoperative troponin elevation. For example, if 2 individuals were otherwise identical, more epochs of ST segment depression usually result in less risk because of the concurrent increase in standard deviation. Therefore, if new ST segment depression was the major reason for measuring troponin levels (which we agree is likely), then the study would tend to exclude those at higher risk (those with a small standard deviation). This could lead to an underestimation of the association between intraoperative ST segment values and postoperative troponin elevation. However, selection bias can produce both overestimation and underestimation of risk, and additional prospective studies in which all subjects are screened for troponin elevation are needed to clarify the relationship among ST segment elevation, depression, and variability.
The final point concerns the overall usefulness of intraoperative ST segments for predicting postoperative myocardial injury. It is true that the ST segment measures examined in our study, along with many of the other features commonly used to predict risk for postoperative troponin elevation, did not discriminate extremely well between those with and without postoperative troponin elevation. Despite this, it did provide a small improvement (net reclassification improvement of 0.0345; 95% confidence interval, 0.00016–0.0591, P = .0474). This is similar to many other patient characteristics, which, by themselves, cannot predict adverse events, but when integrated with other clinical variables, can be used to guide our care of patients. Although the suggested strategy of monitoring postoperative troponin levels for all patients older than the age of 45 years who are admitted to the hospital after noncardiac surgery would clearly be more sensitive, algorithms that use pre-existing data to identify high-risk individuals for screening may be more cost-effective. Therefore, we believe our findings support additional studies aimed at finding ways to use intraoperative ST segments to identify patients at risk for myocardial injury. Hopefully, each additional little improvement will move us closer to understanding and reducing the impact of perioperative myocardial injury.
Michael D. Maile, MD, MSMilo C. Engoren, MDKevin K. Tremper, MD, PhDTyler T. Tremper, BSElizabeth S. Jewell, MSSachin Kheterpal, MD, MBADepartment of AnesthesiologyUniversity of MichiganAnn Arbor, Michiganmmaile@med.umich.edu
1. Sessler DI. Does continuous electronic ST-segment monitoring enhance prediction of postoperative troponin elevation? Anesth Analg. 2016;123:1065.
2. Maile MD, Engoren MC, Tremper KK, Tremper TT, Jewell ES, Kheterpal S. Variability of automated intraoperative ST segment values predicts postoperative troponin elevation. Anesth Analg. 2016;122:608615.