Letters to the Editor: Letter to the Editor
To the Editor:
Various baseline characteristics,1 biomarkers,2 and intraoperative blood pressure3 predict myocardial injury after noncardiac surgery (MINS). Maile et al4 asked whether adding automated intraoperative ST-segment analysis to selected clinical factors improves prediction of MINS. The question is reasonable because troponin elevation (with or without symptoms) has a 10% 30-day mortality5 and can promote interventions, including (1) informing patients that they had myocardial injury and are thus at risk for future heart attacks; (2) starting aspirin; (3) initiating statin and/or angiotensin-converting enzyme inhibitor therapy; (4) improving hypertension control; and (5) using a “teachable moment” to encourage lifestyle changes including smoking cessation, sensible diet, and enhanced exercise.6
Curiously, the analysis is presented as a “case-control study design.” Although ST-segment characteristics are presented in patients with and without troponin elevation, the thrust of the analysis is to add ST-segment characteristics to clinical information in an effort to better predict MINS. For example, “the relationships between individual subject characteristics and postoperative troponin elevation were summarized using logistic regression” and “to isolate independent predictors of postoperative troponin elevation, variables from each lead were entered into a nonparsimonious logistic regression predicting postoperative myocardial injury.” The investigators thus used exposure (clinical characteristics and ST-segment abnormalities) to predict outcome (troponin elevation)—which is a retrospective cohort design.7
A limitation the investigators acknowledge is that troponin screening was not routine; instead, the biomarker usually was evaluated in response to symptoms or signs of myocardial injury. A consequence is that only 5.6% of their patients had even a single postoperative troponin measurement, and only 14% of those patients had elevated concentrations. The overall incidence of MINS was thus only 0.8%. For perspective, the multinational incidence among inpatients older than 45 years of age is 8%.5
We were struck by the investigators’ statement that selection bias (in choosing to monitor troponin) “may have led to an underestimation of the association between ST segment depression and postoperative troponin elevation.” To the extent that ST-segment depression increases both the likelihood of troponin monitoring and the risk of MINS—which is surely the case—selection bias will most likely lead to overestimation of the association between ST-segment depression and postoperative troponin elevation.
Adding ST-segment abnormalities and variability to readily available clinical data improved the area under the receiver operating characteristics curve from 0.68 to 0.71—an increase of only 3%. Although highly statistically significant, 3% represents a trivial improvement and offers little support for the investigators’ conclusion that “automated ST-segment monitoring obtained during surgery is possibly useful for the detection of patients at risk for postoperative myocardial injury.” A more reasonable conclusion might be that ST segment analysis is not especially helpful.
Furthermore, an area under the curve of only about 0.7 indicates that discrimination was poor—with or without ST-segment analysis. Even assuming excellent calibration (not reported), clinicians thus cannot reliably predict which patients will experience postoperative myocardial injury. And precisely because MINS cannot be reliably predicted, troponin screening should be routine for most patients older than 45 years of age having inpatient surgery.6
In summary, it was possible and perhaps even probable that automated intraoperative ST-segment analysis would substantially enhance clinicians’ ability to predict myocardial injury after noncardiac surgery. Maile et al4 present an elegant analysis showing that adding ST-segment analysis only trivially improves discrimination compared with a model based on clinical characteristics alone. The presented results thus appear to directly contradict the authors’ conclusion that “automated ST segment values obtained during anesthesia may be useful for improving the prediction of postoperative troponin elevation.” In fact, their results show that ST-segment analysis adds little value.
Daniel I. Sessler, MD
Department of Outcomes Research
1. The Vascular events In noncardiac Surgery patIents cOhort evaluatioN (VISION) Investigators. Myocardial injury after noncardiac surgery: a large, international, prospective cohort study establishing diagnostic criteria, characteristics, predictors, and 30-day outcomes. Anesthesiology. 2014;120:564–578.
2. Rodseth RN, Biccard BM, Le Manach Y, et al. The prognostic value of pre-operative and post-operative B-type natriuretic peptides in patients undergoing noncardiac surgery: B-type natriuretic peptide and N-terminal fragment of pro-B-type natriuretic peptide: a systematic review and individual patient data meta-analysis. J Am Coll Cardiol. 2014;63:170–180.
3. Mascha EJ, Yang D, Weiss S, Sessler DI. Intraoperative mean arterial pressure variability and 30-day mortality in patients having noncardiac surgery. Anesthesiology. 2015;123:79–91.
4. 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;122:608–615.
5. The Vascular Events In Noncardiac Surgery Patients Cohort Evaluation (VISION) Study Investigators: Association between postoperative troponin levels and 30-day mortality among patients undergoing noncardiac surgery. JAMA. 2012;307:2295–304.
6. Devereaux PJ, Sessler DI. Cardiac complications in patients undergoing major noncardiac surgery. N Engl J Med. 2015;373:2258–2269.
7. Sessler DI, Imrey PB. Clinical research methodology 2: observational clinical research. Anesth Analg. 2015;121:1043–1051.