High-sensitivity cardiac troponin T in young, healthy adults undergoing non-cardiac surgery

Abstract

Background

It is unclear if isolated postoperative cardiac-troponin elevation, often referred to as myocardial injury, represents a pathological event, as control studies in otherwise healthy adults are lacking.

Methods

In this single-centre prospective observational cohort study, serial high-sensitivity cardiac troponin T (hscTnT) plasma concentrations were obtained from young, healthy adults undergoing elective orthopaedic surgery at three time points: before operation, 2–6 h, and 18–30 h after surgery. End points were hscTnT increases after surgery: ≥20% (exceeding analytical variability), ≥50% (exceeding short-term biological variability), and ≥85% (exceeding long-term biological variability). The secondary end point was myocardial injury, defined as new postoperative hscTnT elevation >99th % upper reference limit (URL) (women >10 ng litre⁻¹; men >15 ng litre⁻¹).

Results

Amongst the study population (n=95), no hscTnT increase ≥20% was detected in 68 patients (73%). A hscTnT increase between 20% and 49% was observed in 17 patients (18%), 50–84% in seven patients (7%), and ≥85% in three patients (3%). Twenty patients (21%) had an absolute ΔhscTnT between 0 and 2 ng litre⁻¹, 12 patients (13%) between 2 and 4 ng litre⁻¹, three patients between 4 and 6 ng litre⁻¹, and one patient (1%) between 6 and 8 ng litre⁻¹. Myocardial injury (new hscTnT elevation >99th%) was diagnosed in one patient (1%). The median hscTnT concentrations did not increase after operation, and were 4 (3.9–5, inter-quartile range) ng litre⁻¹ at baseline, 4 (3.9–5) ng litre⁻¹ at 2–6 h after surgery, and 4 (3.9–5) ng litre⁻¹ on postoperative day 1.

Conclusions

One in four young adult patients without known cardiovascular disease developed a postoperative hscTnT increase, but without exceeding the 99th% URL and without evidence of myocardial ischaemia. These results may have important ramifications for the concept of postoperative myocardial injury, as they suggest that, in some patients, postoperative cardiac-troponin increases may be the result of a normal physiological process in the surgical setting.

Clinical trial registration

NCT 02394288.

Editor's key points.

• •Troponin I and T are regulatory proteins necessary for muscle contraction, and are sensitive and specific indicators of myocardial damage.
• •There are many non-ischaemic causes of troponin elevation.
• •High-sensitivity troponin assays have lower cut-off values to indicate ‘abnormal’.
• •This study suggests that some postoperative troponin elevations can be expected in young, healthy patients.

Myocardial injury after non-cardiac surgery, defined as an isolated postoperative cardiac troponin (cTn) elevation without evidence for myocardial ischaemia, is common, particularly amongst patients with or at risk for cardiovascular disease. Even minor postoperative cTn elevations, seen in 11–24%1, 2, 3 of patients with increased cardiovascular risk, are prognostically important, and are associated with increased morbidity and mortality in these patients.1, 2, 3, 4, 5, 6 However, elevated cTn is not by itself evidence of an acute ischaemic event, and there are multiple cardiac-related mechanisms for cTn elevations, including myocardial ischaemia, infarction, pulmonary embolism, heart failure, myocarditis, or—in a trauma population—chest and cardiac trauma. There are also non-cardiac causes for elevated cTn, the most prominent of which is renal failure.⁷

It is also clear that cTn may be released during normal physiological stress. For example, a recent study in healthy triathletes found a >300% increase of cTn concentrations after 1 h of high-intensity exercise.⁸ A study involving healthy volunteers showed that a brief dobutamine infusion causes an increase in high-sensitivity cardiac troponin T (hscTnT).⁹ Likewise, rapid atrial pacing leads to a significant increase in hscTnT, both in patients with or without significant coronary-artery disease.¹⁰ It is, therefore, conceivable that postoperative cTn elevation may occur in patients without cardiovascular risk factors.

The aim of this study, therefore, was to determine if young, healthy patients devoid of known cardiac disease would develop postoperative increases in hscTnT after non-cardiac surgery. High-sensitivity cTn assays have a substantially increased sensitivity and are able to measure small changes in cTn values.7, 11, 12, 13 Our expectation was that we would not observe a hscTnT increase after surgery.

Methods

Design and setting

We conducted a prospective cohort study of patients who underwent elective orthopaedic surgery on the upper or lower extremity under general or regional anaesthesia. Perioperative treatment was at the discretion of the clinical team and was not influenced by study participation. The study was approved by the Institutional Review Board of the Medical University of Vienna and registered at ClinicalTrials.gov (NCT02394288). Informed consent was obtained from each patient. Recruitment was done from patients scheduled for elective orthopaedic surgery at the Vienna General Hospital, Vienna, Austria, from March 2015 to March 2016. Reporting followed the recommendations from the STROBE initiative.¹⁴

Study population

Eligible patients were 18–35 yr old, scheduled for elective orthopaedic surgery and had ASA physical status I or II. Patients were excluded from participation if they had a history or symptoms of cardiac disease, kidney disease, pulmonary embolism, thrombosis, stroke, diabetes, and head or chest trauma, or were pregnant.

Measurements

Patient characteristics, medical history, home medication, and type of surgery were recorded. Vital parameters, type and duration of anaesthesia, tourniquet use, and fluid balance were captured through the electronic anaesthesia record. Serial blood samples and 12-lead ECGs were obtained at three time points for each patient: before operation on the day of surgery, 2–6 h [postoperative day (POD) 0], and 18–30 h (POD 1) after surgery.

For the assessment of hscTnT plasma concentrations, 3 ml of whole blood was drawn into K₃EDTA-coated polyethylene terephthalate tubes (Greiner Bio-One, Kremsmünster, Austria) at each reported time point, and transported within 30 min to the Biobank (www.biobank.at), a centralized sample processing and storage facility at the Department of Laboratory Medicine, Medical University of Vienna. The samples were transferred to an automated pre-analytical system (Roche Diagnostics, Rotkreuz, Switzerland) for further centrifugation (1884 × g, 10 min, ambient temperature) and aliquotation into 2D barcoded polypropylene tubes. Aliquots were then stored until analysis at –70°C. After thawing, hscTnT was quantified by means of electro-chemiluminescence immunoassays using CE-labelled Troponin T hs STAT kits (Roche Diagnostics) on a cobas e 602 integrated into a cobas 8000 modular analyser series (Roche Diagnostics). All samples were analysed in the same batch/run to minimize analytical variability. For this kit/test system combination, the manufacturer reports an intra-assay precision of 2.5–3.7% coefficient of variability for concentrations between 10 and 17 ng litre⁻¹. ECGs were checked for signs of myocardial ischaemia according to the criteria of the Third Universal Definition of Myocardial Infarction¹⁵ by a blinded expert investigator. At each sampling time point, the patients were checked for evidence of adverse cardiac or pulmonary events and fever by a study team member.

Outcomes

End points were selected based on the analytical and biological variability of hscTnT. Postoperative hscTnT change values of ≥20%, ≥50%, and ≥85% were probed. The first value was chosen based on the analytical variability of the hscTnT assay; the ≥50% and ≥85% hscTnT increase from short-term biological variability studies.¹⁶ Frequencies of undetectable values [<limit of detection¹⁷ (LOD) is 4 ng litre⁻¹], measurable (≥4 ng litre⁻¹), and hscTnT values exceeding the sex-specific 99th percentile of the upper reference limit (URL) in a healthy population (99th URL) (female: 10 ng litre⁻¹; male: 15 ng litre⁻¹) were also determined.¹⁸ A combination of postoperative elevation >99th URL and perioperative increase ≥85% of hscTnT values was defined as myocardial injury after surgery, based on the most conservative approach using the long-term biological variability of the hscTnT assay.¹⁶

Statistical analysis

To detect an increase of 20% (±50% standard deviation) within patients at a two-sided α-level of 0.05 and a power of 0.95, a sample size of 84 patients was calculated (G*Power version 3.1.9.2, Kiel, Germany). A cohort of 120 patients was enrolled to account for dropouts and missing plasma samples. Only patients with a complete panel of three hscTnT plasma samples available were included for statistical analysis (final sample size, n=95). Values lower than the LOD of 4 ng litre⁻¹ were entered as 3.9 ng litre⁻¹ to prevent overestimation of within-patient change. Because distributions of hscTnT values and its within-patient change were skewed, the Wilcoxon signed-rank test was used to test for a within-patient change in hscTnT, and Friedman test was used to test for changes between preoperative, 2–6 h, and 18–30 h (POD 1) hscTnT values.

Results

The final sample size consisted of 95 patients (Fig. 1). The study cohort characteristics are shown in Table 1. Surgical procedures included knee and shoulder arthroscopy, hardware removal, and open reduction and internal fixation.

Fig 1.

Flow diagram that shows the process of enrollment and exclusion. POD 0, plasma sample obtained 2–6 h after surgery on POD 0; POD 1, plasma sample obtained 18–30 h after surgery on POD 1. POD, postoperative day.

Table 1.

Patient characteristics and perioperative data. hscTnT, high-sensitivity cardiac troponin T; IQR, inter-quartile range; MAP, mean arterial pressure; n.a., not applicable; RPP, rate pressure product

Count, n (%)	Total 95 (100)	Relative hscTnT increase
		<20%	20–49%	50–84%	>85%
		68 (72)	17 (18)	7 (7)	3 (3)
Patient characteristics
Male, n (%)	78 (82)	54 (79)	16 (94)	6 (86)	2 (67)
Caucasian, n (%)	93 (98)	66 (97)	17 (100)	7 (100)	3 (100)
Age (yr), median [IQR]	25 [22–29]	25 [22–30]	24 [21–29]	22 [20–27]	29 [27–n.a.]
Weight (kg), median [IQR]	77 [68–87]	79 [68–87]	75 [69–85]	69 [63–85]	77 [75–n.a.]
Height (cm), median [IQR]	178 [171–184]	178 [171–185]	178 [173–185]	178 [170–179]	174 [150–n.a.]
BMI, median [IQR]	23.7 [22.7–26.8]	23.7 [22.8–27.2]	23.7 [21.9–26.1]	22.1 [21.5–26.8]	24.9 [24.8–n.a.]
Current smoker, n (%)	40 (43)	26 (38)	8 (47)	3 (43)	3 (100)
Former smoker, n (%)	15 (16)	12 (18)	2 (12)	1 (14)	0 (0)
Pack years, median [IQR]	5 [2–10]	7 [2–10]	2 [1–4]	6 [1–19]	6 [5–n.a.]
ASA status I, n (%)	87 (92)	61 (90)	16 (94)	7 (100)	3 (100)
Hypercholesterolaemia, yes/no/unknown, n (%)	8 (8)/64 (67)/23 (24)	5 (7)/44 (65)/19 (28)	2 (12)/12 (71)/3 (18)	1 (14)/6 (86)/0 (0)	0 (0)/2 (67)/1 (33)
Positive family history, n (%)	14 (15)	10 (15)	1 (6)	2 (29)	1 (33)
Perioperative data
Type of surgery, arm/leg, n (%)	30 (32)/65 (68)	22 (32)/46 (68)	6 (35)/11 (65)	2 (29)/5 (71)	0 (0)/3 (100)
Duration of surgery (min), median [IQR]	64 [40–103]	60 [40–95]	100 [67–125]	23 [15–85]	103 [44–n.a.]
Tourniquet, yes (%)	26 (27)	18 (26)	7 (41)	1 (14)	0 (0)
Type of anaesthesia, general/regional/combined, n (%)	73 (77)/15 (16)/7 (7)	51 (75)/12 (18)/5 (7)	13 (77)/2 (12)/2 (12)	6 (86)/1 (14)/0 (0)	3 (100)/0 (0)/0 (0)
MAP (mm Hg), median [IQR]	84 [78–94]	83 [78–94]	88 [82–95]	87 [79–94]	81 [65–n.a.]
Systole (mm Hg), median [IQR]	115 [107–127]	113 [106–126]	120 [110–128]	117 [115–131]	113 [91–n.a.]
Heart rate, median [IQR]	68 [60–75]	69 [62–75]	68 [58–78]	61 [59–69]	70 [59–n.a.]
RPP, median [IQR]	7903 [6770–8786]	7741 [6786–8756]	8563 [6916–9231]	7802 [7101–8189]	6733 [6456–n.a.]
Fluid balance, median [IQR]	1200 [1000–1500]	1200 [1000–1500]	1350 [1000–1550]	1350 [1000–1700]	1650 [1190–1650]

The incidence of undetectable (hscTnT <LOD of 4 ng litre⁻¹), measurable (hscTnT 4 ng litre⁻¹ – 99th sex-specific URL), and elevated hscTnT concentrations (hscTnT >99th sex-specific URL) is shown in Table 2.

Table 2.

Frequencies of undetectable, measurable, and elevated hscTnT concentrations. In 30 of 95 (32%) patients, the hscTnT concentrations were undetectable before and after operation. The 99th URL is 10 ng litre⁻¹ for women and 15 ng litre⁻¹ for men. hscTnT, high-sensitivity cardiac troponin T; n, count; URL, upper reference limit.

	Preoperative	Postoperative
Undetectable (<4 ng litre−1), n (%)	43 (45)	36 (38)
Measurable (≥4 ng litre−1 to ≤99th URL), n (%)	51 (54)	57 (60)
Elevated (>99th URL), n (%)	1 (1)	2 (2)

Figure 2 depicts the individual relative and absolute hscTnT values between pre- and observed peak postoperative values. No hscTnT increase ≥20% was detected in 68 patients (72%); a hscTnT increase between 20% and 49% was observed in 17 patients (18%), 50–84% in seven patients (7%), and ≥85% in three patients (3%). Amongst the 68 patients who did not experience an increase in hscTnT, 17 patients (18%) actually had a decrease in hscTnT values, which ranged from –3% to –57% as relative change, and from –0.1 to –5.1 ng litre⁻¹. Figure 3 shows the trend of hscTnT values across the three time points.

Fig 2.

Box plots with Tukey's whiskers of the (a) relative and (c) absolute perioperative peak increase (median and lower quartile coincide in both box plots). (b) and (d) The relative and absolute increase of each individual patient is plotted as a line. cTnT, cardiac troponin T; POD, postoperative day; preOP, before surgery.

Fig 3.

Individual hscTnT concentrations at observed time points. hscTnT, high-sensitivity cardiac troponin T.

Amongst the patients with any hscTnT increase after surgery, 20 patients (21%) had an absolute ΔhscTnT between 0 and 2 ng litre⁻¹, 12 patients (13%) had a ΔhscTnT between 2 and 4 ng litre⁻¹, three patients a ΔhscTnT between 4 and 6 ng litre⁻¹, and one patient (1%) a ΔhscTnT between 6 and 8 ng litre⁻¹.

On average, hscTnT plasma concentrations {median [inter-quartile range (IQR)]} did not increase after surgery, and were 4 (3.9–5) ng litre⁻¹ at baseline, 4 (3.9–5) ng litre⁻¹ 2–6 h after surgery, and 4 (3.9–5) ng litre⁻¹ 18–30 h after surgery (POD 1) (Fig. 4).

Fig 4.

Box plots with Tukey's whiskers of preoperative and postoperative hscTnT concentrations. There was no significant perioperative increase of hscTnT concentrations in young, healthy adults undergoing non-cardiac surgery (Friedman anova: P=0.1). Red dots show outliers above the sex-specific 99th URL. PREOP, preoperative hscTnT concentrations; POD 0, hscTnT concentrations on POD 0; POD 1, hscTnT concentrations on POD 1. anova, analysis of variance; hscTnT, high-sensitivity cardiac troponin T; POD, postoperative day.

The incidence rate of myocardial injury after surgery, defined as a combination of hscTnT elevation >99th sex-specific URL and a ≥85% increase, was 1/95 (1%). Table 3 lists the timing of peak hscTnT value amongst patients who had a ≥20% increase. The majority of postoperative hscTnT increases were observed on POD 1.

Table 3.

Timing of peak hscTnT increase. hscTnT, high-sensitivity cardiac troponin T; n, count; POD, postoperative day.

Relative hscTnT increase	Total	2–6 h after surgery	POD 1
20–49%, n (%)	17	6 (35)	11 (65)
50–84%, n (%)	7	2 (29)	5 (71)
≥85%, n (%)	3	2 (67)	1 (33)
Total	27	10 (37)	17 (63)

ECG observations and adverse events

Two (2%) patients showed new postoperative ECG signs consistent with myocardial ischaemia without clinical symptoms or increase of hscTnT. The incidence of adverse events during the study period was 6/95 (6%), and included palpitations (n=2), dizziness (n=2), and chest pain (n=2). All adverse events resolved in less than 1 h, hscTnT was ≤5 ng litre⁻¹, and no ECG signs of ischaemia were observed.

Discussion

The goal of this investigation was to determine the incidence of postoperative hscTnT increases in a population with a low probability of adverse cardiac events or myocardial injury after non-cardiac surgery. Our study population consisted of young, healthy adults; the hscTnT values did not increase. However, the unexpected finding of this investigation was that a significant subset of patients had small to modest postoperative hscTnT increases. About one in four patients had a relative postoperative hscTnT increase of ≥20% compared with baseline, and one in 10 a hscTnT increase of ≥50%, with a minority (one in 32) exceeding ≥85%. Whilst relative changes appear substantial because of the low values at baseline, the absolute hscTnT increases were small, and in most patients, the resulting hscTnT increase remained <4 ng litre⁻¹.

Potential causes for postoperative cTn elevation

cTn assays have a nearly 100% cardiac tissue specificity, except for some patients with chronic muscular disease who have increased concentrations of circulating hscTnT because of re-expression of presumably fetal proteins.19, 20, 21 It is unlikely, therefore, that skeletal muscle injury during orthopaedic surgery caused postoperative cTn increases in our patient population. We cannot exclude the possibility that there was re-expression of hscTnT in some of the patients from their earlier injury, but we would argue that is unlikely to explain all of the cases. Thus, it is more likely that the overwhelming majority of observed hscTnT increases were the result of damage to the cardiac myocyte cell membrane. cTn release can be caused by multiple mechanisms, including ischaemia, inflammation, or non-ischaemic cell necrosis.¹⁵ In our patient population, the likelihood for myocardial ischaemia was low, as it consisted of young, healthy patients without a history of cardiac disease. Furthermore, only a small subset of patients developed minor ECG changes, and therefore, we cannot rule out that some patients indeed developed myocardial ischaemia leading to the postoperative hscTn increase. We cannot exclude the possibility that some patients might have had other silent clinical conditions that could have caused a hscTnT increase, such as pulmonary embolism.

However, perioperative surgical stress may be the more likely cause for the hscTnT increase observed in some of our patients. An interesting observation into the nature of hscTn elevations stems from cardiac biomarker studies in young, healthy adults during exercise, where a substantial number of athletes develop significant post-exercise hscTn elevation >300%.8, 22 The origin is not well understood, but some data suggest that the aetiology for such elevations is the right ventricle.23, 24 Typically, hscTn elevations after myocardial injury are more sustained (>12 h), whereas exercise-induced hscTn elevation in athletes without cardiac symptoms can be short lived (<12 h),²⁵ although studies are lacking that measure the duration until hscTn values return to baseline. This pattern of early and short-lived increase of hscTn may originate from transiently increased membrane permeability with leakage of cytosolic hscTn without cell death, although this is a controversial issue. Lastly, a mechanism similar to stress-induced cardiomyopathy can lead to substantial myocardial injury and cTn elevation, even in the absence of coronary artery disease, and some patients in our patient population may have experienced substantial amounts of surgical stress leading to myocardial injury. However, hscTn elevations after exercise are presumably not harmful in athletes without cardiovascular disease or symptoms, but this is an area of active investigation.²³ It is conceivable that these increases may be part of normal physiology, and simply be an extension of the physiology seen in response to dobutamine or rapid atrial pacing. Lastly, postoperative cTn elevations may also occur in the presence of acute and chronic kidney disease and injury. Whilst a recent study showed that the origin of cTnT elevations in chronic kidney disease is overwhelmingly cardiac and not impaired renal elimination,²⁶ this is not clear in acute kidney injury, where impaired renal elimination may be the dominant cause.

Clinical implications

It is reassuring that very few of the hscTnT increases we observed rose to above the 99th% URL, and so they should not be confused with acute myocardial infarction (MI), which requires a rising pattern of cTn values that exceed the 99th% URL.¹⁵ As indicated earlier, this may be part of normal physiology. It is also unlikely that a postoperative hscTnT increase of 20%, especially from low preoperative values, has clinical significance in itself. Thus, additional research is needed to determine how much change in hscTnT can be expected after operation as part of normal physiology vs a truly pathological process. This is of particular importance given the modest changes we observed in our putatively normal patients, with most hscTnT increases at <4 ng litre⁻¹. In many situations, changes as modest as 5 ng litre⁻¹ are used to rule in acute myocardial injury and MI.²⁷ This value was also evaluated in the VISION 2 study and was considered an increase. Our data suggest caution with that approach.

The VISION 2 study recently showed a concentration-dependent increased risk of 30-day mortality amongst adult patients older than 45 yr of age who developed an increased postoperative hscTnT value.²⁸ Few patients had symptoms or signs of ischaemia, and thus, some of those increases may well have been attributable to non-ischaemic causes. What was clear that new postoperative hscTnT elevation without ischaemic or overt non-cardiac cause was associated with increased risk; the larger the degree of change, the greater the risk. Our data indicate that some of the smaller increases may be attributable to a physiological response to the surgical procedure, and not be indicative of acute myocardial injury.

Limitations

The patients were not followed for three PODs as in most studies of perioperative myocardial injury. We, therefore, may have missed delayed hscTn elevation. Patients in our cohort were often discharged on POD 1, and we were unable to obtain follow-up information. However, we believe that the risk of late onset hscTn elevation in our population was low. Because the hscTnT assay fails to detect values in some normal patients (compared with high-sensitivity cardiac troponin I (hscTnI) assays), many baseline hscTnT values were considered undetectable resulting in a possible underestimation of the frequency with which hscTnT values change. Although the likelihood of acute kidney injury was very low amongst our patient population, we cannot rule out that changes in kidney function may have influenced postoperative hscTnT values. Lastly, in some patients, we observed an unexpected decrease in postoperative hscTnT concentrations. We can only speculate about the mechanism for such decrease, and believe that either sampling error (drawing samples from an i.v. line) or biological variability may be the most likely cause.

In conclusion, our study showed that postoperative hscTnT elevations occur in young, presumably healthy adults undergoing non-cardiac surgery. These results may have important ramifications for the concept of postoperative myocardial injury. Subsequent research is needed to determine the underlying mechanism of cTn elevation and how to best interpret them after non-cardiac surgery.

Authors' contributions

Study design: A.D., T.S., P.N.

Data acquisition: A.D., C.W., M.T., M.M., M.H., V.B.W., T.S.

Data analysis: A.D., H.H., T.S., M.M., M.G.S., A.S.J., P.N.

Drafting of the manuscript: A.D., P.N.

Critical revision for important intellectual content: all authors.

Approval of the final version of the manuscript: all authors.

Editorial decision: September 24, 2017

Handling editor: P.S. Myles

Declaration of interest

P.N. has received research grants and other research support from Roche Diagnostics (Indianapolis, IN, USA), and research grants and other research support from Abbott. M.G.S. has received research support from Siemens Healthcare Diagnostics, Abbott Diagnostics, and Instrumentation Laboratory; consultation fees from Instrumentation Laboratory, Becton Dickinson, and Alere; and speaker fees from Abbott. A.S.J. has received consultation fees from Beckman, Abbott, Alere, Critical Diagnostics, Roche, sphingotec, Siemens, Novartis, and theheart.org. The rest of the authors have no conflict of interest.

Funding

National Institutes of Health/National Heart, Lung, and Blood Institute (R01HL126892) to P.N.