Author + information
- Received August 14, 1998
- Revision received January 6, 1999
- Accepted January 21, 1999
- Published online May 1, 1999.
- Franz F Immer, MD∗,
- Franco Stocker, MD†,* (, )
- Andrea M Seiler, MD†,
- Jean-Pierre Pfammatter, MD†,
- Denis Bachmann, MD‡,
- Gert Printzen, MD§ and
- Thierry Carrel, MD∗
- ↵*Reprint requests and correspondence: Dr. Franco Stocker, MD, Division of Pediatric Cardiology, University Hospital, 3010 Berne, Switzerland
It was the aim of the study to test the prognostic value of cardiac troponin-I (cTnI) concerning the early postoperative course after pediatric cardiac surgery.
Cardiac troponin-I is a very specific and sensitive marker of myocardial damage in adults and children. As perioperative myocardial damage may be a significant factor of postoperative cardiac performance, serial cTnI values were analyzed in children undergoing open heart surgery.
Seventy-three children undergoing elective correction of congenital heart disease including atrial and ventricular surgical manipulation were studied. Cardiac troponin-I levels were measured serially and correlated with intra- and postoperative parameters (such as doses and length of inotropic support, renal and hepatic function, duration of intubation). Patients with prolonged postoperative recovery were analyzed with special attention to the cTnI levels.
The cutoff point for the definition of a high and a low risk group of cTnI values was set at 25 μg/liter, 4 h after admission to the intensive care unit (ICU) and at 35 μg/liter considering the maximal value of cTnI in the first 24 h in the ICU. The results showed a highly significant correlation between the need for inotropic support, the severity of renal dysfunction and the duration of intubation in relation to the serum levels of cTnI.
Cardiac troponin-I serum levels after open heart surgery in children and infants 4 h after admission to the ICU allowed anticipation of the postoperative course and correlated with the incidence of significant postoperative complications.
Troponin-I (cTnI), as a part of the tropomyosin complex, is a well established marker of myocardial damage in adults (1). In children and infants, we and others have demonstrated that determination of cTnI serum levels allows a very specific assessment of myocardial damage (2–4).
In several adult studies that have included patients with unstable angina pectoris (5,6)and critically ill intensive care unit (ICU) patients (7), pathologic values of cTnI have been shown to have a prognostic value regarding late morbidity and mortality.
The aim of the present study was to evaluate the prognostic value of serum levels of cTnI after open heart surgery in children and to correlate cTnI levels with the early postoperative period.
Seventy-six infants and children with congenital heart disease, scheduled for cardiac surgery with ventricular surgical manipulation, were evaluated. Seventy-three of them fulfilled the inclusion criteria: elective operation for ventricular septal defect (VSD), atrioventricular (AV) defect or tetralogy of Fallot and serial blood samples of cTnI up to 24 h after admission to the ICU. The children were prospectively enrolled between March 1995 and May 1998.
The mean age of these 73 patients, corresponding to 100% of the study population, was 21 ± 28 months (range 10 days to 133 months), and 39 were male. Closure of a VSD was performed in 31, repair of a tetralogy of Fallot or double outlet right ventricle in 19 and in 23 patients an AV defect was repaired (20 had a complete and 3 an incomplete AV defect). All patients were operated using moderate hypothermia (24 to 30°C) and cardiopulmonary bypass, myocardial protection was realized with antegrade cold blood cardioplegia. Ventricular manipulation was either done by transatrial approach or through a ventriculotomy. Three patients were excluded; two of them died in the early postoperative period and had no complete blood samples of cTnI up to 24 h after admission to the ICU, and one patient was excluded after a successful repair of a tetralogy of Fallot because he had to be cannulated on an emergency basis due to a severe cyanotic episode during induction of anesthesia. The study protocol was approved by the University Ethical Committee.
Blood samples were collected preoperatively as baseline and in the postoperative course immediately after admission to the ICU and then at 4, 8, 12 and 18–24 h. As the ethical committee allowed only blood samples at the time of other routine blood tests, it was not possible to have complete series of cTnI determination for all patients. For the measurements of cTnI we used a Sandwich-Immunoassay-Test (Dade-Behringer, Miami, Florida). All the analyses were performed at 37°C. According to this protocol the upper limit of normal for cTnI was set at 0.6 μg/liter (4).
The following peri- and postoperative variables were all taken into account and correlated with cTnI levels: cardiopulmonary bypass time (CPB), aortic cross-clamping time and reperfusion time, duration of intubation, doses and length of inotropic support by catecholamines, renal and hepatic function, as well as potentially severe postoperative complications such as life-threatening arrhythmias, low cardiac output and renal failure requiring peritoneal dialysis or exitus; these variables were extracted from the intensive care reports. All these factors were analyzed and correlated with the cTnI values.
In a first step the maximum cTnI values during the first 24 h after admission to the ICU were analyzed. The patients were divided into two groups. The cutoff point for cTnI was set at 35 μg/liter, on the basis of the known kinetic of cTnI and preliminary statistical analysis. According to their maximal postoperative cTnI values during the first 24 h, 54 patients were included in the first group (low risk) with maximal cTnI values lower than 35 μg/liter, and the remaining 19 patients, with cTnI values exceeding 35 μg/liter, were defined as the second group (high risk).
In a second step we evaluated the predictive value of earlier cTnI measurements: in 30 patients at the admission to the ICU (with a cutoff point at 20 μg/liter) and in 40 patients 4 h later (with a cutoff point at 25 μg/liter).
In addition we analyzed the predictive value of duration of cardiopulmonary bypass and age of the patients.
Statistical tests were performed by using a two-tailed ztest for average differences of a standard repartition. A p value less than 0.05 was considered statistically significant. Results are expressed as mean values ± standard deviation. Finally the positive (PPV) and negative predictive value (NPV) were calculated.
Maximal troponin-I values in the first 24 hours (n = 73 patients)
Out of the total of 73 patients, we found in 54 patients (mean age 24 ± 31 months, range 10 days to 133 months), corresponding to the lower risk group (group 1), cTnI values lower than 35 μg/liter in the first 24 h after admission to the ICU. Looking at the distribution we found 26 patients with a closure of a VSD, 14 patients with a repair of tetralogy of Fallot and 14 patients with a correction of an AV defect. Group 2 is formed by 19 patients (mean age 16 ± 28 months, range 2 months to 125 months) in whom cTnI values exceeded 35 μg/liter. Two patients died in the postoperative period—both of them were in group 2, with an increase of the maximal values for cTnI of 126 and 319 μg/liter. The intraoperative parameters and the necessity for vasoactive support, duration of intubation as well as renal and hepatic function are shown in Table 1. A statistically significant difference was reached in all parameters. Especially renal dysfunction, reflected by serum creatinine and serum urea and vasoactive support, showed a highly significant difference between group 1 and 2. Looking at the 15 patients with potentially severe postoperative complications (defined under Methods), all of them are included in group 2; 8 of them had transient renal failure with peritoneal dialysis, 4 had severe postoperative low cardiac output, 2 patients died from a multiorgan failure and one had a life-threatening arrhythmia requiring defibrillation. Considering these results we found for the prediction of severe postoperative complications a PPV of 100% and a NPV of 93% in cTnI values exceeding 35 μg/liter, during the first 24 h in the ICU (Fig. 1).
Troponin-I values 4 hours after admission to the ICU (n = 40 patients)
In accordance with the known kinetics of cTnI and statistical analysis the cutoff point was set at 25 μg/liter, 4 h after admission to the ICU. In this time period 29 patients (mean age 19 ± 23 months, range 10 days to 88 months) showed cTnI values below 25 μg/liter and in 11 patients (mean age 28 ± 41 months, range 4 to 125 months) cTnI exceeded 25 μg/liter. For these two groups the necessity for vasoactive support, duration of intubation as well as the incidence of renal and hepatic dysfunction are summarized in Table 1. The 2 patients who died in the postoperative period as well as 8 of the 9 patients with severe postoperative complications were, 4 h after admission to the ICU, in group 2 according to their high cTnI levels, resulting in a PPV of 89% and a NPV of 94% for the anticipation of severe postoperative complications (Fig. 1).
Troponin-I values at admission to the ICU (n = 30 patients)
In 30 patients we had cTnI values at admission to the ICU. Looking at the kinetics of cTnI and statistical analysis the cutoff point between low and high risk groups, was set at 20 μg/liter. Twenty-four children (mean age 21 ± 30 months, range 10 days to 125 months) were thus assigned to the first group (low risk group) with cTnI values below 20 μg/liter and six patients (mean age 12 ± 3 months, range 9 to 17 months) to the second group (high risk group) exceeding 20 μg/liter. The results of these patients are displayed in Table 1. Six patients with blood samples at the admission to the ICU fulfilled the criteria for severe postoperative complications in accordance with the intensive care report. Only three of them were classified as high risk patients at the admission, with cTnI values exceeding 20 μg/liter. Considering these results obtained with patients with severe postoperative complications the PPV came to 50% and the NPV to 89% (Fig. 1).
To assess the effect of duration of cardiopulmonary bypass and age of the patients we analyzed the results for the two parameters in the total collective.
Duration of cardiopulmonary bypass
The cutoff point of CPB was set at 100 min for the definition of a high (≥100 min) and a low (<100 min) risk group. Forty-five patients (mean age 20 ± 27 months, range 10 days to 122 months) had CPB below 100 min, corresponding to group 1, compared with group 2 with 28 patients (mean age 25 ± 35 months, range 21 days to 125 months) with CPB exceeding 100 min. Statistical significance was reached for cTnI values at 4 h as well as looking at the maximal values of cTnI in the first 24 h postoperatively (Table 2). The doses of vasoactive support, serum values of renal and hepatic function and duration of intubation also reached statistical significance. The PPV for the assessment of severe postoperative complications in relation to the duration of CPB comes to 80% and the NPV is set at 72%.
Of the overall collective of 73 patients, 45 patients are infants below 1 year of age (mean age 6 ± 3 months, range 10 days to 12 months), compared with 28 patients older than 12 months (mean age 25 ± 35 months, range 13 to 133 months). As displayed in Table 2, no statistical significance was reached for the parameters of the CPB, cTnI values, vasoactive support, duration of intubation or renal and hepatic function.
As already shown in previous studies, cTnI represents an important marker in the assessment of myocardial damage in children and infants after pediatric cardiac surgery (3,4,8). Troponin-I is a more sensitive marker of myocardial cell damage than creatine kinase MB fraction in the postoperative pediatric population (4,8). It has been shown that at least in children the absolute values of cTnI correlate very well with the extent of perioperative myocardial damage (4). The possibility of predicting the postoperative outcome in children is very attractive, allowing anticipation of the need for vasoactive support and duration of intubation, as well as having a prognostic aspect.
Preliminary statistical analysis of our results were performed to determine a cutoff point of cTnI values for defining a low and a high risk group of patients at the time of admission, 4 h after admission to the ICU and considering the maximal cTnI values during the first 24 h after admission to the ICU. By that we tried to predict the postoperative course and outcome in correlation with the cTnI values.
Four hours after admission and during the first 24 h to the ICU the difference between the two groups (regarding vasoactive support, duration of intubation and renal and hepatic dysfunction) are statistically significant (Table 1). Furthermore severe postoperative complications correlated well with cTnI values 4 h after the entrance to the ICU and come even up to a PPV of 100% in the first 24 h after admission to the ICU. The two patients dying in the postoperative period both had cTnI values exceeding 100 μg/liter, underlining Taggart’s (8)finding that in children with very high postoperative cTnI values (>100 μg/liter) the outcome is lethal, (9).
On the other hand the results found immediately after admission to the ICU are less reliable, but a NPV of 89% (Fig. 1)represents important information for the medical staff on the ICU in regard to the expectation of an uneventful postoperative course.
A prolonged duration of CPB leads to more pronounced myocardial damage reflected by higher cTnI values, which has already been shown in adults (10). A tendency to higher cTnI values is also described by Hirsch et al. in a recently published study (3), which confirms our results. But CPB alone does not have the same impact for a prediction of the postoperative course as cTnI.
A tendency to higher cTnI values in children below 1 year of age is also discussed in published reports (3,8). However, due to the small number of patients and to the wide spectrum of surgical interventions, until now no statistical analysis has been performed (8). In our study, looking at similar surgical interventions, no statistical significance was found between cTnI in children below 1 year of age compared with older children.
In conclusion, cTnI seems to be a very promising marker in the prediction of early postoperative course in infants and children undergoing cardiac surgery. The maximal cTnI value in the first 24 h, as well as the results 4 h after the admission to the ICU, allow a reproducible prediction of severe postoperative complications and of the duration and the extent of intensive care treatment.
- cardiopulmonary bypass time
- cardiac troponin-I
- intensive care unit
- negative predictive value
- positive predictive value
- ventricular septal defect
- Received August 14, 1998.
- Revision received January 6, 1999.
- Accepted January 21, 1999.
- American College of Cardiology
- Galvani M,
- Ottani M,
- Ferrini D,
- et al.
- Lüscher M.S,
- Thygesen K,
- Ravkilde J,
- Heickendorff L
- Mair J,
- Larue C,
- Mair P,
- Balogh D,
- Calzolari C,
- Puschendorf B