Intensive Care Medicine

, Volume 29, Issue 12, pp 2317–2322

Elevated troponin T concentrations in critically ill patients


    • Department of Clinical ChemistryCanisius-Wilhelmina Hospital
  • Joris J. J. P. M. van de Leur
    • Department of Internal Medicine and Intensive Care MedicineCanisius-Wilhelmina Hospital

DOI: 10.1007/s00134-003-1953-2

Cite this article as:
Klein Gunnewiek, J.M.T. & van de Leur, J.J.J.P.M. Intensive Care Med (2003) 29: 2317. doi:10.1007/s00134-003-1953-2



To determine the incidence of troponin T elevations among a selected group of critically ill patients, to correlate these findings to electrocardiographs, and to compare troponin T-positive and T-negative patients in relation to clinical parameters.


Prospective study.


Mixed surgical and medical intensive care unit.


Thirty-four consecutive critically ill patients who were mechanically ventilated or underwent thoracic or vascular surgery.


Blood samples were collected at admission, the next morning, and 24 h after the second blood sampling. These samples were used for troponin T measurement and electrocardiographs were made when troponin T levels were elevated.

Main results

Eleven of 34 patients (32%) had elevated troponin T levels, which were already present upon admission in eight out of 11 patients (73%). Most patients underwent surgery prior to ICU admission (21 of 34 patients). Significantly (P=0.0055) more troponin T-positive patients underwent acute surgery, and significantly more (P=0.045) troponin T-positive patients suffered from hypotension. Only four of the troponin T-positive patients were diagnosed as suffering from an acute myocardial infarction based on electrocardiographs. All troponin T-positive patients had coronary artery disease: nine had a history of CAD and two had actual CAD. No difference in mortality rates was observed between troponin T-positive and T-negative patients.


An unexpectedly high percentage of included patients had troponin T elevations, which could be corroborated by electrocardiographs in only four cases suggesting that a high percentage of critically ill patients with a history of CAD suffer from clinically unrecognised (minor) myocardial damage.


Troponin TCritically illHypotension


Damage of myocardial cells results in the release of contractile-regulating proteins, such as cardiac troponin T and I (cTnT and cTnI, respectively), into the circulation. Since these cardiac troponins differ from troponins present in skeletal muscles, and that cardiac troponins are normally not found in blood, cTnT and cTnI are highly specific markers for myocardial cell damage [1, 2, 3, 4]. Elevated cTnT and cTnI levels are not only observed in acute myocardial infarction (AMI) [5, 6] but also in unstable angina [1, 7], myocarditis [8, 9], cardiac trauma [10], and perioperative cardiac complications [11, 12, 13]. In addition, increased cTnI [14, 15, 16, 17, 19] and cTnT [16, 18] concentrations have been detected in blood from critically ill patients with sepsis or septic shock. In these patients with sepsis, troponin elevations (ranging from 36% to 85% of included patients) were associated with clinically unrecognised myocardial cell injury and an increased mortality rate. Elevated cTnI levels have also been detected in critically ill patients suffering from hypotension [19]. Huge differences in percentages of elevated cardiac troponin levels were observed in these studies and different troponin assays were used.

We studied thirty-four critically ill patients admitted to the ICU. Inclusion criteria were mechanical ventilation, and thoracic or vascular surgery. Cardiac TnT levels were measured in order to determine the incidence of cTnT elevations among this group of patients in our hospital.

Patients and methods


During a 2-month period, we included 34 consecutive patients (mean age 64 years, range 24–79 years; male-to-female ratio, 23/11), who were admitted to our intensive care unit (ICU). Inclusion criteria were mechanical ventilation, thoracic (no cardiac) surgery or vascular surgery. Patients with underlying cardiac disease or chronic renal failure were not excluded from this study. Blood samples were collected at admission, the next morning, and 24 h after the second blood sample was taken. In some patients the third blood sampling was not performed until they had already left the ICU [seven out of 23 (30%) patients without cTnT elevation and 1 out of 11 (9%) patients with cTnT elevation]. The three collected blood samples were used for cTnT, CK, ASAT, ALAT, LDH, urea, and creatinine measurements. When cTnT was elevated (>0.1 μg/l) electrocardiographs were made. The Acute Physiology And Chronic Health Evaluation (APACHE) score was determined upon ICU admission.

Laboratory parameters

Cardiac troponin T concentration was measured using the troponin T STAT immunoassay on an Elecsys 2010 (Roche). This third generation test uses two monoclonal antibodies specifically directed against human cardiac troponin T. Although the new cut-off for the third-generation troponin T test is reduced to 0.035 μg/l, taking into account the analytical variability and individual biological changes, we used the recommended 0.1 μg/l as the clinical threshold value based on physicians’ experience and the fact that most of our included patients did not have the classical signs or symptoms of AMI. Creatine kinase [upper reference limit: 200 U/l (male) and 170 U/l (female)], ALAT [upper reference limit: 50 U/l (male) and 40 U/l (female)], ASAT (upper reference limit: 40 U/l), LDH (upper reference limit: 450 U/l), urea (upper reference limit: 7 mmol/l), and creatinine [upper reference limit: 120 μmol/l (male) and 100 μmol/l (female)] were measured on a Modular analyser (Roche).

Electrocardiographic studies

A 12-lead electrocardiograph (ECG) was performed when cTnT levels were elevated (>0.1 μg/l).

Statistical analysis

The Fisher exact (two-sided) and the Student’s t-test (two-sided) were used to compare variables. For both tests a level of P<0.05 was considered significant.


Thirty-four patients were included in this study. Using the cut-off of 0.1 μg/l, 23 (68%) of these included patients did not have elevated cTnT concentrations while 11 (32%) had. Eight patients (73%) already had cTnT elevations upon entry in this study while only three cTnT-positive patients had initially normal cTnT (<0.1 μg/l) concentrations. Table 1 shows the characteristics of the included cTnT-positive and T-negative patients. Overall cTnT-positive patients were older than cTnT-negative patients were (P=0.006). In addition, nine cTnT-positive patients had a history of coronary artery disease (CAD; myocardial infarction, angina pectoris or CABG) and the other two cTnT-positive patients were diagnosed for CAD just before or at ICU admission. In comparison, only 30% of cTnT-negative patients had a history of CAD (P=0.0001). Four patients had cTnT levels between 0.035 and 0.1 μg/l. These patients did not have a history of CAD and suffered from sepsis (two patients) or shock (one patient).
Table 1.

Characteristics of cTnT-positive and -negative patients

High cTnT (>0.1 μg/l) n=11

Low cTnT (<0.1 μg/l) =23


Mean age

73 years (66–79)

59 years (24–76)


Male-to-female ratio



History of CAD

9/11 (82%)

7/23 (30%)


CAD (history or actual)

11/11 (100%)

7/23 (30%)



7 (64%)

14 (61%)


5/7 (71%)

1/14 (7%)



2/7 (29%)

13/14 (93%)


1/7 (14%)

7/14 (50%)


2/7 (29%)

3/14 (21%)


4/7 (57%)

3/14 (21%)


1/14 (8%)


4 (36%)

9 (39%)


3/4 (75%)

4/9 (45%)



3/9 (33%)


1/4 (25%)

2/9 (22%)

Mechanical ventilation

8/11 (73%)

13/23 (57%)

Death inside ICU

2 (18%)

4 (17%)

Death outside ICU

2 (18%)

0 (0%)

Stay at ICUa

5 days (2–10 days)

6.2 days (1–34 days)

*P<0.05; **P<0.001

aDeceased patients included

The included patients were divided into two groups: those patients undergoing surgery prior to ICU admission and patients admitted to ICU based on medical grounds (Table 1). Surgical patients were further classified as acute and non-acute surgical patients. Seven out of 11 cTnT-positive patients underwent surgery (five acute surgeries) and 14 out of 23 cTnT-negative patients (one acute surgery) underwent surgery. Among cTnT-positive patients one lobectomy, two (re)laparotomies, and four vascular surgeries (two aneurysm aortae abdominalis and two femoropopliteal bypasses) were performed. In addition, three cTnT-positive patients suffered from pneumonia and one suffered from a myocardial infarction and ventricular fibrillation. Among cTnT-negative patients, seven lobectomies, three laparotomies, three vascular surgeries (all elective aneurysm aortae abdominalis), and one thymectomy were performed. Four cTnT-negative patients suffered from pneumonia, three suffered from sepsis (due to a carcinoma, arthritis, and an abscess) and two patients suffered from subarachnoid haemorrhage.

Four characteristics were significantly different between cTnT-positive and T-negative patients: 1) all cTnT-positive patients had CAD compared to 30% of the cTnT-negative patients (P=0.0001); 2) mean age of cTnT-positive patients was higher (P=0.006) when compared to cTnT-negative patients (73 versus 59 years, respectively); 3) more (P=0.045) cTnT-positive patients suffered from hypotension (systolic pressure below 90 mmHg, Table 2); and 4) a significantly higher percentage (P=0.0055) of cTnT-positive surgical patients underwent acute surgery when compared to cTnT-negative patients (71% versus 7%, respectively). Thoracic surgery seemed to be more prevalent among cTnT-negative patients while vascular surgery was more prevalent among cTnT-positive patients. The need for inotropics during the first 3 days seemed to be higher among cTnT-positive patients (73%) compared to cTnT-negative patients (35%). Among cTnT-positive patients, there seems to be a relation between the extent of cTnT elevation and hypotension. Furthermore, cTnT elevation showed a small non-significant increase in mortality rate. No correlation between cTnT and creatinine kinase (CK) was observed (Table 2).
Table 2.

Clinical and biochemical parameters of cTnT-positive and -negative patients. (HR heart rate, BP blood pressure)

High cTnT (>0.1 μg/l) n=11

Low cTnT (<0.1 μg/l) n=23


Elevated CK levels

8 (73%)

16 (70%)

Mean APACHE score



Use of inotropics

8/11 (73%)

8/23 (35%)

Systolic hypotensiona

6/11 (55%)

4/23 (17%)


Mean HR (beats per min)




aSystolic hypotension: <90 mmHg

Analysis of electrocardiographic patterns of ten out of 11 cTnT-positive patients using the redefined criteria for myocardial infarction [20] showed no significant changes among three patients (Table 3). All three patients suffered from hypotension. Four ECG patterns were conclusive for myocardial infarction (ST elevation or depression). Two of these four patients were clinically diagnosed as having AMI while the other two patients suffered from hypotension and hypotension in combination with tachycardia. Three ECGs were non-specific (only negative T waves). These three patients suffered from atrial fibrillation (one patient), atrial fibrillation in combination with hypotension (one patient), and congestive heart failure (one patient). No correlation between ECG results and the need for mechanical ventilation was observed. Nearly all patients suffering from hypotension (9/10) needed inotropics.
Table 3.

cTnT elevations in relation to electrocardiography

High cTnT (>0.1 μg/l)

Myocardial infarction


Non-specific (negative T waves)


No electrocardiographic changes


No electrocardiography available


The lowest cTnT value reported in cTnT-positive patients was 0.12 μg/l while the highest reported value was 2.16 μg/l. There was no correlation between cTnT elevation and ECG results or medical history.


In order to determine the incidence of cTnT elevations among critically ill patients in our hospital we studied a group of 34 consecutive ICU patients who were mechanically ventilated or subjected to either thoracic or vascular surgery. Eleven patients (32%) had cTnT elevations. Only four of these cTnT-positive patients were recognised as suffering from a myocardial infarction based on electrocardiographs, indicating that a high percentage of our critically ill patients suffer from unnoticed myocardial damage. As many as eight cTnT-positive patients (73%) already had cTnT elevations upon admission to the ICU. Twenty-one of all included patients underwent surgery prior to ICU submission. A significantly higher percentage (P=0.0055) of cTnT-positive surgical patients underwent acute surgery when compared to cTnT-negative patients (5/7 cTnT-positive patients versus 1/14 cTnT-negative patients, see Table 1).

Other studies describe elevated cardiac troponin concentrations, cTnT as well as cTnI, among different groups of critically ill patients. Guest and co-workers [21] as well as Kollef [22] and co-workers measured cTnI levels in 209 and 260 critically ill patients, respectively. All patients admitted to the ICU were included and no selection prior to inclusion was performed. In these two studies, 15% [21] and 15.8% [22] of the included patients were cTnI-positive. Cardiac TnT and/or TnI elevations have also been studied in selected groups of critically ill patients with sepsis [14, 15, 16, 17, 18, 19]. The reported incidences of cardiac troponin elevations among these critically ill patients ranged from 50% to 85% for cTnI [14, 15, 16, 17, 19] (19, 15, 46, 20, and 31 patients were included, respectively) and from 36% to 69% for cTnT [16, 18] (46 and 26 patients were included, respectively). In the present study, we found 32% cTnT-positivity. The reported differences in cardiac troponin-positivity can be the consequence of different inclusion criteria used to select patients and the use of different cardiac troponin assays. Ver Elst and co-workers [16] compared cTnI with cTnT and reported that all cTnI-negative patients were also cTnT-negative, but seven out of 46 patients with sepsis who were cTnI-positive at some time point remained cTnT-negative.

All of our cTnT-positive patients had a history of CAD or actual CAD. This percentage was significantly lower (30%) among cTnT-negative patients (P=0.0001). Arlati and co-workers [19] also found that seven of 31 included patients had a history of CAD and all seven had abnormal cTnI levels. In addition, significantly more (P=0.0055) of our cTnT-positive patients underwent acute surgery in comparison to cTnT-negative patients (71% versus 7%). These results suggest that patients with pre-existing CAD are at risk of developing new myocardial damage when they are (acute) critically ill. This is also indicated by our finding that a significant correlation (P=0.045) between cTnT-positivity and hypotension exists. Fifty-five percent of cTnT-positive patients and 17% of cTnT-negative patients were hypotensive. This has also been shown by others [19, 21]. Arlati and co-workers [19] found cTnT-positivity among all included hypotensive patients and the median concentration of cTnI increased with the duration of hypotension. These results indicate that blood pressure is a very important determinant for maintenance of the myocardial flow. Hypotension can lead to myocardial damage, especially in critically ill patients with underlying CAD.

In this study, cTnT-positivity is significantly related to patients’ age (P=0.006), and a weak association between cTnT elevations and hospital mortality was observed (mortality among 4/11 cTnT-positive and 4/23 cTnT-negative patients). The sample size in this study was too small (34 patients) to investigate whether cTnT is an independent marker of hospital mortality. In the two studies that included larger numbers of critically ill patients (260 and 219 patients) [21, 22], cTnI-positive patients had a greater hospital mortality rate, but elevated cTnI concentrations did not independently contribute to the prediction of hospital mortality. We found a correlation between cTnT-positivity and the need for mechanical ventilation (73% and 57% for cTnT-positive and T-negative patients, respectively) and inotropics (73% and 35% for cTnT-positive and T-negative patients, respectively). We found no correlation between cTnT concentrations and CK levels. This could be expected since CK lacks specificity in the setting of skeletal muscle disease or injury, which frequently occurs among critically ill patients and patients who underwent surgery.

Analysis of ECG patterns of our cTnT-positive patients (see Table 3) shows that four ECG patterns were conclusive for myocardial infarction, three patterns showed no changes, and three patterns were non-specific. All three patients without ECG changes suffered from hypotension and needed inotropic support. These results suggest that cTnT elevation among these patients is caused by reduction of oxygen supply as a consequence of hypotension that possibly results in diffuse cardiac damage that cannot be detected by ECG. In addition, one of these patients was treated with hemodialysis. Elevated cTnT values are detectable in about 30% of patients suffering from renal failure (e.g., chronic hemodialysis patients) without acute coronary disease [23, 24]. However, this patient did have normal cTnT concentrations at the time of entry in the study, whereas cTnT levels were elevated 2 days after inclusion suggesting that the observed cTnT elevation is due to new cardiac damage.

In studies performed among ICU patients with sepsis [14, 16, 18], no electrocardiographic changes whatsoever were observed among cardiac troponin-positive patients. Hence, troponin elevations were subscribed to clinically unrecognised myocardial cell injuries. The autopsy of a small group of cTnI-positive patients excluded the presence of myocardial infarction using microscopic examination although contraction band necrosis, an early marker of irreversible myocyte injury, tended to be more frequent in cTnI-positive patients [16]. These results and ours show that (minimal) cardiac damage is a common feature among critically ill patients, especially when these patients suffer from pre-existing CAD. This (minimal) cardiac damage might be caused by a temporary reduction in oxygen supply (e.g., micro-occlusions, hypotension, and anaemia) in combination with increased myocardial oxygen consumption or increasing wall stress (e.g., congestive heart failure), hypertension with left ventricular hypertrophy, tachycardia, pulmonary embolism or myocardial toxins (as a consequence of sepsis) (reviewed by [4]). Wu [25] hypothesised that in some cases reversible injury might be present. This reversible damage could be the result of increased permeability of myocytes to macromolecules caused by the release of myocardial depressant substances such as TNF. This might be true in patients with sepsis with a rapid rise and fall of troponin levels. However, in our study we did not observe a rapid rise and fall of cTnT, and all cTnT-positive patients had a history of CAD making them vulnerable to additional myocardial damage when they are critically ill. Therefore, our elevated cTnT levels are a consequence of irreversible rather than reversible injury.

What lessons can we learn from these and other data obtained among ICU patients? Should cTnT measurements be performed more often among these patients? Furthermore, if cTnT levels are elevated in the absence of electrocardiographic changes, is it wise to treat cTnT-positive patients differently (e.g., other drugs)? Our finding that a high percentage of critically ill patients with a history of CAD in combination with acute surgery, hypotension, atrial fibrillation or tachycardia have clinically unrecognised myocardial damage suggests that troponin T is a valuable tool for diagnosis. It can be difficult to establish the exact time of onset of cardiac damage since 73% of our included patients were already cTnT-positive upon admission to the ICU and elevated troponin levels can persist (cTnT can remain elevated for up to 14 days). This might hamper formulation of therapeutic strategies. Jaffe and colleagues recommend in their editorial [4] that when a cause for troponin increase is detected, therapy should be predicated on this cause. If no cause can be ascertained, clinical follow-up is all that clinicians can do. We have shown here that the detection of the cause for a cTnT increase among our patients was made possible by careful observation of the clinical course in combination with the patient’s history.

In conclusion, an unexpectedly high percentage of critically ill patients (32%) admitted at our ICU had elevated cTnT concentrations. All cTnT-positive patients had a history of CAD and only four cTnT-positive patients were diagnosed with myocardial infarction. Our results indicate that patients with a history of CAD are at risk of developing cardiac damage in critical clinical situations (e.g., acute surgery, hypotension, tachycardia, and atrial fibrillation) which may go unnoticed. Troponin T measurement can be used to identify these patients.


We would like to thank Dr. E. Lamferts for critical evaluation of the electrocardiographs.

Copyright information

© Springer-Verlag 2003