Secondary Logo

Journal Logo

Patient Safety: Research Report

Timing of Preoperative Troponin Elevations and Postoperative Mortality After Noncardiac Surgery

Maile, Michael D. MD, MS; Jewell, Elizabeth S. MS; Engoren, Milo C. MD

Author Information
doi: 10.1213/ANE.0000000000001309


Inadequate coronary perfusion can lead to cellular damage with subsequent release of intracellular components into the circulation. Some of these proteins, such as cardiac troponin (cTn), have isoforms that differ from those found in other muscles.1 Because of this specificity, cTn is a cornerstone in the diagnosis of acute coronary syndrome (ACS).2 Furthermore, the magnitude of cTn elevation in ACS is correlated with the risk of short-term mortality.3 Given the importance of early diagnosis and treatment, cTn concentrations are measured liberally in at-risk patients.

With improvements in the sensitivity of clinically used assays, it is increasingly recognized that many patients have low, but detectable, levels of serum cTn outside the setting of ACS.4 Elevations of cTn indicative of cardiac strain or injury, but not associated with coronary atherosclerosis, include such diverse causes as perimyocarditis, endocarditis, takotsubo cardiomyopathy, radiofrequency catheter ablation, cardiac contusion, strenuous exercise, and sympathomimetic drugs.2,4 Although sometimes trivialized as only a troponin leak, several studies have demonstrated that, although not produced by myocardial ischemia or ACS, these elevations are associated with adverse outcomes.5–7 Although preoperative myocardial infarction is an established risk factor for adverse events,8 less is known about patients with preoperative cTn elevation not related to ACS. This creates a precarious situation for clinicians, given that not only are optimal management strategies undefined, but also the risk of proceeding to surgery is poorly quantified. For example, although a preoperative cTnT level >14 ng/L was associated with increased postoperative mortality,9 we do not know whether the postoperative mortality risk is proportional to the magnitude of cTn elevation, if there is a risk for smaller elevations, or if this risk decreases with longer time between peak cTn measurement and time of surgery.

To answer these questions we conducted a retrospective cohort study of patients undergoing noncardiac surgery who underwent cTnI testing before surgery. In addition to analyzing the effect of concentration, the time between peak cTnI levels and surgery was also examined. We hypothesized that both high cTnI levels and short duration between the peak level and surgery were associated with increased risk of postoperative mortality.


Study Population

This retrospective study was approved by the IRB, which waived informed consent because it only involved deidentified, previously collected data. Patients at least 18 years old who had a cTnI measurement made within 30 days before a noncardiac surgical procedure at the University of Michigan were included (Fig. 1). Current procedural terminology codes were used to identify the individuals who had undergone a preoperative percutaneous coronary intervention, and this group was excluded from the analysis. The time period included was March 1, 2006 to June 5, 2013. The start date was chosen as the first date all data were reliably available in our electronic medical record and the last date was chosen as the last date before submission to the IRB. Given the retrospective nature of this research, measurement of preoperative cTnI concentration was at the discretion of the managing service and was based on clinical reasons. Typical indications include chest pain, hemodynamic instability, and electrocardiographic abnormalities. At our institution, cTnI (Troponin I Ultra assay; Siemens Healthcare Diagnostics, Deerfield, IL) is measured and values as low as 0.10 ng/mL are reported. The upper limit for normal, or reference range, is <0.30 ng/mL and was constant throughout the study period. Of note, although this assay can detect lower troponin levels, the hospital laboratory does not report them. Instead, they are reported as undetected.

Figure 1.
Figure 1.:
Diagram describing the derivation of the study population. The initial list of eligible subjects was found by searching for noncardiac surgical cases in which the individual had a troponin level measured during the 30 days preceding the operation.

Subjects with detected cTnI concentrations were divided into 9 cohorts using terciles of cTnI levels and terciles of time between this measurement and surgery. Patients with undetected cTnI levels served as controls. Emergent surgical procedures were removed from the analysis and patients who underwent a nonemergent, noncardiac surgical procedure and had their preoperative troponin level measured during the 30 preceding days were included in the analysis. The following subject characteristics were also collected: age, sex, ASA physical status, surgical duration, and several laboratory values (bicarbonate, creatinine, total bilirubin, lactate, and brain natriuretic peptide). Missing data were present for laboratory values and body mass index. These were imputed, with 10 iterations per imputation, using IVEware version 2.0 (Institute for Social Research, University of Michigan, Ann Arbor, MI) and SAS software, Version 9.3 for Windows (SAS Institute, Cary, NC).


The primary outcome of this study was 30-day postoperative mortality. These data were obtained from hospital records and the Michigan death index.


Data were described using means and SDs for normally distributed characteristics and median and interquartile range for nonnormally distributed variables. Study cohorts were created by dividing subjects with detectable preoperative cTnI levels into 9 groups according to time terciles: short (0–2.10 days), medium (>2.10–8.86 days), and long (>8.86–30 days) and cTnI terciles: low (0.10–0.23 ng/mL), medium (0.23–0.84 ng/mL), and high (0.84–494.20 ng/mL) (Fig. 2). This method of categorization was used given the lack of data to define what represents a clinically significant preoperative troponin level. Subjects with undetectable preoperative cTnI levels comprised the tenth (reference) group.

Figure 2.
Figure 2.:
Scatterplot demonstrating the distribution of preoperative troponin concentrations and duration of time between the highest measured troponin value and surgery for the study population with detectable preoperative troponin levels. Lines reflect the 9 study cohorts, which were generated by dividing subjects into troponin terciles and time terciles. The study population is skewed toward low troponin concentrations that were measured in shortly before surgery.
Table 1.
Table 1.:
Preoperative Subject Characteristics

The Cochran-Armitage trend test was used to determine whether there was a change in risk among the different cohorts. Bivariable and multivariable logistic regressions were used to compare the incidence of 30-day mortality among groups and to adjust for the differences in risk factors. All preoperative patient information (Table 1) was included in the multivariable models, regardless of their univariable association with 30-day mortality. The odds ratio (OR) and 95% confidence interval (CI) were used to describe the magnitude and precision of associations. For all statistical tests, a P value <0.05 was used to denote statistical significance. Statistical analysis was completed using R Core Team (2013) (R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL:


During the study period, 4575 patients had cTnI levels drawn within 30 days before a noncardiac surgical procedure. They were predominantly middle aged and had a variety of comorbidities (Table 1). Of these patients, 986 (21.6%) had cTnI >0.10 ng/mL (Fig. 2) and 281 patients (6.1%) died within 30 days. The mortality for those with undetectable and detectable preoperative troponin levels was 4.7% and 11.4%, respectively. Any increase in cTnI above the reported level (0.10 ng/dL) was associated with increased 30-day mortality. This risk increased from 4.7% in the undetectable cTnI group to 12.7% in high tercile group. Compared with those with undetectable levels, the OR of 30-day mortality progressively increased from 1.98 (95% CI, 1.33–2.95) in the low cTnI tercile to 2.95 (95% CI, 2.06–4.22) in the high tercile.

For patients in the low cTnI tercile, having surgery <2.10 days or between 2.10 and 8.86 days after the cTnI measurement was associated with increased 30-day mortality. However, if surgery occurred >8.86 days after cTnI elevation, the risk was similar to patients without elevation (Fig. 3). For patients who had more extensive cTnI elevations (>0.23 ng/mL), the unadjusted OR of 30-day mortality was higher than in the control group, except for the medium-troponin/short-time group, which had an OR similar to the control group (OR = 1.53; 95% CI, 0.55–4.27) (Fig. 3). Examination of this risk across time terciles revealed a trend between lower mortality and longer time between the peak cTnI concentration and surgery for the low cTnI cohort (P = 0.0238) but not the medium (P = 0.09) or high (P = 0.20) cohorts.

Figure 3.
Figure 3.:
Heatmap summarizing the relationship between time and troponin concentration and 30-day mortality. Text in each box provides the odds ratio, 95% confidence interval, and the proportion of individuals experiencing 30-day mortality. Risk appears to remain elevated longer for those subjects who had a higher peak troponin level. Risk decreases with longer amounts of time between the peak troponin level and surgery (P = 0.0032).
Table 2.
Table 2.:
Multivariable Regression
Figure 4.
Figure 4.:
Adjusted heatmap based on multivariable logistic regression that includes preoperative patient characteristics. Text in each box provides the odds ratio, 95% confidence interval, and the proportion of individuals experiencing 30-day mortality. Even small levels of troponin remain associated with increased mortality when the time to surgery is <2.1 days.

After using multivariable logistic regression to adjust for surgical specialty, demographics, comorbidities, and other laboratory values, preoperative cTnI values remained associated with 30-day mortality for 3 of the 9 groups (Table 2 and Fig. 4). To assess interactions between the troponin/time categories and other significant variables, additional logistic regression models were developed with these interaction terms as a sensitivity analysis (Supplemental Digital Content, Although a few interactions were statistically significant for individual troponin/time categories, none was consistent across either troponin concentration or time. In addition, likelihood ratio tests between the models including and not including the interactions were not significant. Therefore, the final model did not include these interaction terms.


Although cTn elevation is most frequently thought of as a biomarker of ACS, studies have demonstrated its importance outside of this setting. Terms, such as myocardial injury, troponin leak, and secondary unstable angina, are all used to describe the situation in which these biomarkers are released from the heart in the absence of coronary artery blockage.10 This situation has been documented in multiple settings (Table 3), such as sepsis,11 pulmonary embolism,12 renal failure,13 and stroke,14 and these non-ACS cTn elevations are associated with increased risk for mortality even in the general population.6,7

Table 3.
Table 3.:
Noncoronary Causes of Troponin Elevation2 , 4 , 11–14

Although some of our patients may have had classical myocardial infarction, more than one-third (the entire low tercile and some of the medium tercile of troponin values) were classified as having normal troponin values by our laboratory, yet they were at increased risk of 30-day mortality. This suggests that our laboratory criteria for normal troponin values in this population need to be revised. Study is also needed to determine whether troponin values below our limit of detection are associated with increased risk.

This study expands our knowledge of the impact of cTn levels on surgical risk for those undergoing noncardiac surgery. Weber et al.9 previously demonstrated that, in a high-risk population undergoing noncardiac surgery, high-sensitivity cTnT values provide additional prognostic information to the revised cardiac risk index. Although they examined cTnT rather than cTnI, their study found an association between small changes in cTnT concentrations and mortality, with levels >14 ng/L associated with a 160% (27%–431%) increase in the death hazard. Our study extends their results by demonstrating a dose-response relationship between the amount of time that elapses between a small peak troponin value and survival, especially in those with lower troponin levels. In other words, the increased risk of death is partially ameliorated by a longer wait from peak cTnI to surgery for those with small increases in preoperative troponin levels. Until now, the effect of time between peak cTnI levels and surgery was unstudied. Although our retrospective study cannot establish a causal relationship, it suggests that waiting a longer time after cTn elevation before surgery might reduce 30-day mortality in certain patients. We have no explanation why the patients with an intermediate cTnI elevation demonstrated a time-to-surgery trend in the opposite direction from the small and large cTnI elevation groups. This study raises the question of why a similar trend was not seen in all of the troponin concentration groups. Possible explanations include that these represent a different disease process with different characteristics, that longer periods of time are necessary for larger troponin elevations, or that these patients are being managed differently.

Until now, the effect of time between peak cTnI levels and surgery was unstudied. Although our retrospective study cannot establish a causal relationship, it suggests that waiting a longer time after cTn elevation before surgery might reduce 30-day mortality. Unfortunately, our study does not provide any insight on preoperative management, such as β-blockers and antiplatelet agents, or intraoperative management, such as different arterial blood pressure and heart rate goals that might modify outcome in these patients. Obtaining a better understanding of the pathophysiology of cTn release and how it relates to myocardial injury and death will lead to strategies to reduce its occurrence and association with mortality in the surgery population.15

Although we found that longer time between cTn elevation and surgery was associated with lower risk of death before adjusting for other patient characteristics, additional studies are needed to determine whether surgery should be delayed for individuals with a recent troponin elevation and, if so, to what level troponin values should be decreased. Our results suggest this may be longer than a month given that some individuals remained at increased risk in our multivariable model (details pertaining to this model are provided as Supplemental Digital Content, Some guidance may be inferred from studies of myocardial injury after radiofrequency ablation. In this setting, higher levels of cTn were associated with greater inflammation that persisted long after resolution of cTn levels, which was associated with greater risk of recurrence.16 Given that patients remain at elevated risk for 3 months after an ablation procedure,16–18 perhaps this would be a reasonable time to wait after myocardial injury before proceeding with a surgical procedure until future studies bring more clarity to this situation.

Our findings should be interpreted in light of several limitations. First, this study was retrospective, and, therefore, factors not measured may confound the results. For example, given the high mortality rate of those both with and without detectable troponin levels, our study population is clearly skewed toward those with a high baseline risk of mortality. Based on this, these findings may not apply to individuals who clinicians do not feel are at risk of myocardial ischemia. However, since serum troponins are not typically used as a screening test before noncardiac surgery, this selection bias does not negate the findings of this study. Second, differences in clinical practice among institutions may limit the ability to generalize our findings to locations that are more or less likely to measure cTn. Third, our laboratory only reports concentrations >0.10 ng/mL; therefore, these results do not provide information about values lower than this level. We also do not know the reasons for the surgery and for the timing of the surgery. Although we adjusted for surgical specialty, we do not know the specific surgeries, and it is possible that patients with greater elevations of troponin underwent more complicated operations within each specialty. Emergency surgery may not be able to be delayed and the patients and physicians most consider the elevated risks associated with having an elevated cTn. Finally, because we do not know the either the cause of the cTn elevation or the cause of mortality, we do not know whether the myocardial injury as shown by the elevated cTnI contributed to death or the underlying pathology that led to the myocardial injury, and cTnI elevation was the contributing factor. Further study is needed to clarify this.

The main strength of our study is that, by using an electronic medical record, we were able to analyze a large group of patients, thus giving us sufficient statistical power to determine the relationship of cTnI level and time to surgery on mortality. By excluding subjects who underwent preoperative percutaneous coronary intervention, our findings are most applicable to patients with disease entities involving presumed coronary blood flow-demand imbalance rather than type 1 myocardial infarction.2

In conclusion, we found that the mortality risk is both magnitude and time related. Higher levels of preoperative cTnI were associated with higher postoperative mortality, and longer time to surgery appeared to reduce this risk for individuals with mild preoperative troponin elevations. Prospective studies are needed to determine whether delaying surgery in patients with elevated preoperative troponin levels improves postoperative outcomes.


Name: Michael D. Maile, MD, MS.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Michael D. Maile has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Elizabeth S. Jewell, MS.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Elizabeth S. Jewell has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.

Name: Milo C. Engoren, MD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Milo C. Engoren has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

This manuscript was handled by: Sorin J. Brull, MD.


1. Bodor GS, Porterfield D, Voss EM, Smith S, Apple FSCardiac troponin-I is not expressed in fetal and healthy or diseased adult human skeletal muscle tissue.Clin Chem19954117105
2. Thygesen K, Alpert JS, Jaffe AS, Simoons ML, Chaitman BR, White HD, Katus HA, Lindahl B, Morrow DA, Clemmensen PM, Johanson P, Hod H, Underwood R, Bax JJ, Bonow RO, Pinto F, Gibbons RJ, Fox KA, Atar D, Newby LK, Galvani M, Hamm CW, Uretsky BF, Steg PG, Wijns W, Bassand JP, Menasché P, Ravkilde J, Ohman EM, Antman EM, Wallentin LC, Armstrong PW, Simoons ML, Januzzi JL, Nieminen MS, Gheorghiade M, Filippatos G, Luepker RV, Fortmann SP, Rosamond WD, Levy D, Wood D, Smith SC, Hu D, Lopez-Sendon JL, Robertson RM, Weaver D, Tendera M, Bove AA, Parkhomenko AN, Vasilieva EJ, Mendis SJoint ESC/ACCF/AHA/WHF Task Force for the Universal Definition of Myocardial InfarctionThird universal definition of myocardial infarction.Circulation2012126202035
3. Antman EM, Tanasijevic MJ, Thompson B, Schactman M, McCabe CH, Cannon CP, Fischer GA, Fung AY, Thompson C, Wybenga D, Braunwald ECardiac-specific troponin I levels to predict the risk of mortality in patients with acute coronary syndromes.N Engl J Med199633513429
4. Agewall S, Giannitsis E, Jernberg T, Katus HTroponin elevation in coronary vs. non-coronary disease.Eur Heart J20113240411
5. Bjurman C, Larsson M, Johanson P, Petzold M, Lindahl B, Fu ML, Hammarsten OSmall changes in troponin T levels are common in patients with non-ST-segment elevation myocardial infarction and are linked to higher mortality.J Am Coll Cardiol20136212318
6. de Lemos JA, Drazner MH, Omland T, Ayers CR, Khera A, Rohatgi A, Hashim I, Berry JD, Das SR, Morrow DA, McGuire DKAssociation of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population.JAMA2010304250312
7. Eggers KM, Venge P, Lindahl B, Lind LCardiac troponin I levels measured with a high-sensitive assay increase over time and are strong predictors of mortality in an elderly population.J Am Coll Cardiol201361190613
8. Lee TH, Marcantonio ER, Mangione CM, Thomas EJ, Polanczyk CA, Cook EF, Sugarbaker DJ, Donaldson MC, Poss R, Ho KK, Ludwig LE, Pedan A, Goldman LDerivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery.Circulation199910010439
9. Weber M, Luchner A, Seeberger M, Manfred S, Mueller C, Liebetrau C, Schlitt A, Apostolovic S, Jankovic R, Bankovic D, Jovic M, Mitrovic V, Nef H, Mollmann H, Hamm CWIncremental value of high-sensitive troponin T in addition to the revised cardiac index for peri-operative risk stratification in non-cardiac surgery.Eur Heart J20133485362
10. Pierpont GL, McFalls EOInterpreting troponin elevations: do we need multiple diagnoses?Eur Heart J2009301358
11. Spies C, Haude V, Fitzner R, Schröder K, Overbeck M, Runkel N, Schaffartzik WSerum cardiac troponin T as a prognostic marker in early sepsis.Chest1998113105563
12. Ng AC, Yong AS, Chow V, Chung T, Freedman SB, Kritharides LCardiac troponin-T and the prediction of acute and long-term mortality after acute pulmonary embolism.Int J Cardiol201316512633
13. Dierkes J, Domröse U, Westphal S, Ambrosch A, Bosselmann HP, Neumann KH, Luley CCardiac troponin T predicts mortality in patients with end-stage renal disease.Circulation200010219649
14. Sandhu R, Aronow WS, Rajdev A, Sukhija R, Amin H, D’aquila K, Sangha ARelation of cardiac troponin I levels with in-hospital mortality in patients with ischemic stroke, intracerebral hemorrhage, and subarachnoid hemorrhage.Am J Cardiol20081026324
15. van Kimmenade RR, Januzzi JL JrWhose heart will get broken? Troponin testing and future heart failure.Eur Heart J2009307556
16. McCabe JM, Smith LM, Tseng ZH, Badhwar N, Lee BK, Lee RJ, Scheinman MM, Olgin JE, Marcus GMProtracted CRP elevation after atrial fibrillation ablation.Pacing Clin Electrophysiol200831114651
17. Richter B, Gwechenberger M, Socas A, Zorn G, Albinni S, Marx M, Bergler-Klein J, Binder T, Wojta J, Gössinger HDMarkers of oxidative stress after ablation of atrial fibrillation are associated with inflammation, delivered radiofrequency energy and early recurrence of atrial fibrillation.Clin Res Cardiol201210121725
18. Lim HS, Schultz C, Dang J, Alasady M, Lau DH, Brooks AG, Wong CX, Roberts-Thomson KC, Young GD, Worthley MI, Sanders P, Willoughby SRTime course of inflammation, myocardial injury, and prothrombotic response after radiofrequency catheter ablation for atrial fibrillation.Circ Arrhythm Electrophysiol20147839

Supplemental Digital Content

Copyright © 2016 International Anesthesia Research Society